Systems and methods for navigating a host vehicle

Abstract

In one implementation, a method includes receiving at least one image captured by a camera of a host vehicle from an environment of the host vehicle; analyzing the at least one image to identify a region of interest; selecting a portion of the at least one image based on the region of interest; receiving map information associated with the environment of the host vehicle; providing the portion of the at least one image and the map information to a trained system; and receiving an output provided by the trained system. The output includes identification of an object in the environment of the host vehicle and location information for the object relative to the map information.The method further includes causing the host vehicle to initiate at least one navigation action based on the identification of the object and the location information for the object.

Description

PRIOR ARTCross-references to related applicationsThis application claims priority to U.S. Provisional Applications No. 63 / 618,503, filed January 8, 2024, and No. 63 / 678,210, filed August 1, 2024. The above applications are incorporated herein by reference in their entirety.Background InformationAs technology continues to progress, the goal of a fully autonomous vehicle capable of navigating the roads is becoming in graspable proximity. Autonomous vehicles may need to take a variety of factors into account and make appropriate decisions based on these factors to reach an intended destination safely and accurately. For example, an autonomous vehicle may need to process and interpret visual information (e.g., information captured by a camera) and may also use information from other sources (e.g., from a GPS device, a speed sensor, an accelerometer, a suspension sensor, etc.). At the same time, to navigate to a destination, an autonomous vehicle may also need to identify its location within a particular roadway (e.g., a particular lane within a multi-lane road), navigate amongst other vehicles, avoid obstacles and pedestrians, observe traffic signals and signs, and travel from one road to another road at appropriate intersections or intersections. Utilizing and interpreting large amounts of information collected from an autonomous vehicle as the vehicle travels to its destination presents a variety of design challenges. The low level of data (e.g., captured image data, map data, GPS data, sensor data, etc.) that an autonomous vehicle may need to analyze, retrieve, and / or store poses challenges that may actually limit or even adversely affect autonomous navigation. Moreover, when an autonomous vehicle relies on traditional map technology to navigate, the tighter volume of data needed to store and update the map presents tremendous challenges.BRIEF DESCRIPTION OF THE DRAWINGSThe accompanying drawings, which are incorporated in and constitute a part of this disclosure, illustrate various disclosed embodiments. The following are shown: FIG. 1 is a schematic illustration of an example system, in accordance with the disclosed embodiments. FIG. 2A is a schematic side view illustration of an example vehicle including a system, in accordance with the disclosed embodiments. FIG. 2B is a schematic top view illustration of the vehicle and system shown in FIG. 2A, in accordance with the disclosed embodiments. FIG. 2C is a schematic top view illustration of another embodiment of a vehicle including a system, in accordance with the disclosed embodiments. FIG. 2D is a schematic top view illustration of yet another embodiment of a vehicle including a system, in accordance with the disclosed embodiments. FIG. 2E is a schematic top view illustration of yet another embodiment of a vehicle including a system, in accordance with the disclosed embodiments. FIG. 2F is a schematic illustration of example vehicle control systems, in accordance with the disclosed embodiments. FIG. 3A is a schematic illustration of an interior of a vehicle including a rearview mirror and a user interface for a vehicle imaging system, in accordance with the disclosed embodiments. FIG. 3B is an illustration of an example of a camera mount configured to be positioned behind a rearview mirror and against a vehicle windshield, in accordance with the disclosed embodiments. FIG. 3C illustrates the camera mount shown in FIG. 3B from a different perspective, in accordance with the disclosed embodiments. FIG. 3D is an illustration of an example of a camera mount configured to be positioned behind a rearview mirror and against a vehicle windshield, in accordance with the disclosed embodiments. FIG. 4 is an example block diagram of a memory configured to store instructions for performing one or more operations, in accordance with the disclosed embodiments. FIG. 5A is a flow diagram illustrating an example process for causing one or more navigation responses based on monocular image analysis, in accordance with the disclosed embodiments. FIG. 5B is a flow diagram illustrating an example process for detecting one or more vehicles and / or pedestrians in a set of images, in accordance with the disclosed embodiments. FIG. 5C is a flow diagram illustrating an example process for detecting road markings and / or lane geometry information in a set of images, in accordance with the disclosed embodiments. FIG. 5D is a flow diagram illustrating an example process for detecting traffic lights in a set of images, in accordance with the disclosed embodiments. FIG. 5E is a flow chart illustrating an example process for causing one or more navigation responses based on a vehicle path, in accordance with the disclosed embodiments. FIG. 5F is a flowchart illustrating an example process for determining whether a preceding vehicle is changing lanes, in accordance with the disclosed embodiments. FIG. 6 is a flow chart illustrating an example process for causing one or more navigation responses based on stereo image analysis, in accordance with the disclosed embodiments. FIG. 7 is a flow chart illustrating an example process for causing one or more navigation responses based on analysis of three sets of images, in accordance with the disclosed embodiments. FIG. 8 illustrates a sparse map for providing autonomous vehicle navigation, in accordance with the disclosed embodiments. FIG. 9A is a polynomial representation of portions of a road segment in accordance with the disclosed embodiments. FIG. 9B is a three-dimensional space curve illustrating a target trajectory of a vehicle for a particular road segment included in a sparse map, in accordance with the disclosed embodiments. FIG. 10 illustrates example landmarks that may be included in a sparse map, in accordance with the disclosed embodiments. FIG. 11A illustrates polynomial representations of trajectories, in accordance with the disclosed embodiments. FIGS. 11B and 11C illustrate target trajectories along a multi-lane road, in accordance with the disclosed embodiments. FIG. 11D illustrates an example road signature profile, in accordance with the disclosed embodiments. FIG. 12 is a schematic illustration of a system using crowdsourced data received from a plurality of vehicles for autonomous vehicle navigation, in accordance with the disclosed embodiments. FIG. 13 illustrates an example autonomous vehicle road navigation model represented by a plurality of three-dimensional splines, in accordance with the disclosed embodiments. FIG. 14 illustrates a map skeleton generated by combining location information from multiple trips, in accordance with the disclosed embodiments. FIG. 15 illustrates an example of a longitudinal orientation of two trips with example signs as landmarks, in accordance with the disclosed embodiments. FIG. 16 illustrates an example of a longitudinal orientation of many trips with an example sign as an landmark, in accordance with the disclosed embodiments. FIG. 17 is a schematic illustration of a system for generating travel data using a camera, a vehicle, and a server, in accordance with the disclosed embodiments. FIG. 18 is a schematic illustration of a system for crowdsourced sparse card, in accordance with the disclosed embodiments. FIG. 19 is a flow chart illustrating an example process for generating a sparse map for autonomous vehicle navigation along a road segment, in accordance with the disclosed embodiments. FIG. 20 is a block diagram of a server in accordance with the disclosed embodiments. FIG. 21 is a block diagram of a memory in accordance with the disclosed embodiments. FIG. 22 illustrates a process for clustering vehicle trajectories associated with vehicles in accordance with the disclosed embodiments. FIG. 23 illustrates a navigation system for a vehicle that may be used for autonomous navigation, in accordance with the disclosed embodiments. FIGS. 24A, 24B, 24C, and 24D illustrate example lane markings that may be detected in accordance with the disclosed embodiments. FIG. 24E illustrates example imaged lane markings, in accordance with the disclosed embodiments. FIG. 24F illustrates an example anomaly associated with detecting a lane marker, in accordance with the disclosed embodiments. FIG. 25A illustrates an example image of the environment of a vehicle for navigation based on the imaged lane markings, in accordance with the disclosed embodiments. FIG. 25B illustrates lateral location correction of a vehicle based on mapped lane markings in a road navigation model, in accordance with the disclosed embodiments. FIGS. 25C and 25D provide conceptual representations of a localization technique for locating a host vehicle along a target trajectory using mapped features included in a sparse map. FIG. 26A is a flow diagram illustrating an example process for mapping a lane marker for use in autonomous vehicle navigation, in accordance with the disclosed embodiments. FIG. 26B is a flow diagram illustrating an example process for autonomously navigating a host vehicle along a road segment using mapped lane markings, in accordance with the disclosed embodiments. FIG. 27 is a flow diagram illustrating an example process for navigating a host vehicle, in accordance with disclosed embodiments. FIG. 28 illustrates an example image captured by a camera of a host vehicle from the environment of the host vehicle in accordance with the disclosed embodiments. FIG. 29 is an exemplary functional block diagram of a trained system in accordance with the disclosed embodiments.SUMMARYEmbodiments consistent with the present disclosure provide systems and methods for navigation of a host vehicle. In one embodiment, a system for navigating a host vehicle is disclosed.The system may comprise at least one processor comprising circuitry and a memory, the memory including instructions that, when executed by the circuitry, cause the at least one processor to: receive at least one image captured by a camera of the host vehicle from an environment of the host vehicle, analyze the at least one image to identify an area of interest in the environment of the host vehicle, select a portion of the at least one image based on the area of interest, and receive map information associated with the environment of the host vehicle. The map information may include one or more identifiers of a route of the host vehicle. The instructions may further cause the at least one processor to provide the portion of the at least one image and the map information to a trained system. The trained system may be configured to generate one or more detections using a language model architecture based on the analysis of the portion of the at least one image and the map information. The instructions may further cause at least the processor to receive output provided by the trained system. The output may include an identifier of an object in the environment of the host vehicle and location information for the object relative to the map information. Further, the instructions may cause the host processor to cause the host vehicle to initiate at least one navigation action based on the identifier of the object and the information about the location of the object.In one embodiment, a method for navigating a host vehicle comprises receiving at least one image captured by a camera of the host vehicle from an environment of the host vehicle; analyzing the at least one image to identify a region of interest in the environment of the host vehicle; selecting a portion of the at least one image based on the region of interest; receiving map information associated with the environment of the host vehicle, the map information including one or more identifiers of a route of the host vehicle; providing the portion of the at least one image and the map information to a trained system, the trained system configured to generate one or more detections using a language model architecture based on the analysis of the portion of the at least one image and the map information; receiving an output provided by the trained system, the output including an identifier of an object in the environment of the host vehicle and location information for the object relative to the map information; and causing the host vehicle to initiate at least one navigation action based on the identifier of the object and the location information for the object.In one embodiment, a non-transitory computer readable medium stores program instructions executable by at least one processor to perform a method. The method includes receiving at least one image captured by a camera of the host vehicle from an environment of the host vehicle; analyzing the at least one image to identify a region of interest in the environment of the host vehicle; selecting a portion of the at least one image based on the region of interest; receiving map information associated with the environment of the host vehicle, the map information including one or more identifiers of a route of the host vehicle; providing the portion of the at least one image and the map information to a trained system, the trained system configured to generate one or more detections using a language model architecture based on the analysis of the portion of the at least one image and the map information; receiving an output provided by the trained system, the output including an identifier of an object in the environment of the host vehicle and location information for the object relative to the map information; and causing the host vehicle to initiate at least one navigation action based on the identifier of the object and the location information for the object.In one embodiment, a non-transitory computer readable medium may store program instructions executable by at least one processor to perform a method. The method may include receiving at least one image captured by a camera of the host vehicle from an environment of the host vehicle, analyzing the at least one image to identify a region of interest in the environment of the host vehicle, selecting a portion of the at least one image based on the region of interest, and receiving map information associated with the environment of the host vehicle. The map information may include one or more identifiers of a route of the host vehicle. The method may further include providing the portion of the at least one image and the map information to a trained system. The trained system may be configured to generate one or more detections using a language model architecture based on the analysis of the portion of the at least one image and the map information. The method may further include receiving output provided by the trained system. The output may include an identifier of an object in the environment of the host vehicle and location information for the object relative to the map information. The method may further include causing the host vehicle to initiate at least one navigation action based on the identification of the object and the location of the object.In one embodiment, a non-transitory computer readable medium may store program instructions executable by at least one processor to perform a method. The method may include receiving at least one image captured by a camera of the host vehicle from an environment of the host vehicle and receiving map information associated with the environment of the host vehicle. The map information may include one or more identifiers of a route of the host vehicle. The method may further include providing the at least one image and the map information to a trained system. The trained system may be configured to generate one or more detections using a language model architecture based on the analysis of the at least one image and the map information. The method may further include receiving output provided by the trained system. The received output may include an identification of an object in the environment of the host vehicle and information about the location of the object relative to the map information. The method may further include causing the host vehicle to initiate at least one navigation action based on the identification of the object and the location of the object.The foregoing general description and the following detailed description are exemplary and explanatory only and do not limit the claims.DETAILED DESCRIPTIONThe following detailed description refers to the accompanying drawings. Where possible, the same reference numerals are used in the drawings and the following description to refer to the same or similar parts. While several illustrative embodiments are described herein, modifications, adaptations, and other implementations are possible. For example, substitutions, additions, or modifications may be made to the components illustrated in the drawings, and the illustrative methods described herein may be modified by replacing, rearranging, removing, or adding steps to the disclosed methods. Accordingly, the following detailed description is not limited to the disclosed embodiments and examples. Rather, the proper scope is defined by the appended claims.Autonomous Vehicle (AV) OverviewAs used throughout this disclosure, the term "autonomous vehicle" refers to a vehicle capable of implementing at least one navigation change without driver input. A "navigation change" refers to a change in one or more of steering, braking, or accelerating the vehicle. To be autonomous, a vehicle need not be fully automatic (e.g., full operation without driver or without driver input). Rather, an autonomous vehicle includes those that may be operated under driver control during certain periods and without driver control during other periods. Autonomous vehicles may also include vehicles that control only some aspects of vehicle navigation, such as steering (e.g., to maintain a vehicle heading between limitations of the vehicle's lane), but may leave other aspects to the driver (e.g., braking). In some cases, autonomous vehicles may handle some or all of the aspects of braking, cruise control, and / or steering the vehicle.Since human drivers typically rely on visual cues and observations to control a vehicle, traffic infrastructures are built accordingly, with lane markings, traffic signs, and traffic lights all configured to provide visual information to the drivers. Given these design characteristics of traffic infrastructures, an autonomous vehicle may include a camera and a processing unit that analyzes visual information captured from the environment of the vehicle. The visual information may include, for example, components of the traffic infrastructure (e.g., lane markings, traffic signs, traffic lights, etc.) that may be observed by drivers and other obstacles (e.g., other vehicles, pedestrians, dirt, etc.). Additionally, an autonomous vehicle may also use stored information, such as information that provides a model of the environment of the vehicle when navigated. For example, the vehicle may use GPS data, sensor data (e.g., from an accelerometer, a speed sensor, a suspension sensor, etc.), and / or other map data to provide information regarding its environment as the vehicle travels, and the vehicle (as well as other vehicles) may use the information to locate on the model.In some embodiments in this disclosure, an autonomous vehicle may use information obtained during navigation (e.g., from a camera, a GPS device, an accelerometer, a speed sensor, a suspension sensor, etc.). In other embodiments, an autonomous vehicle may use information obtained from previous navigations through the vehicle (or other vehicles) during navigation. In still other embodiments, an autonomous vehicle may use a combination of information obtained during navigation and information obtained from previous navigations. The following sections provide an overview of a system consistent with the disclosed embodiments, followed by an overview of a forward imaging system and methods consistent with the system. The sections that follow disclose systems and methods for constructing, using, and updating a sparse map for autonomous vehicle navigation.System OverviewFIG. 1 is a block diagram representation of a system 100, in accordance with the example disclosed embodiments. The system 100 may include various components depending on the requirements of a particular implementation. In some embodiments, the system 100 may include a processing unit 110, an image capturing unit 120, a position sensor 130, one or more storage units 140, 150, a map database 160, a user interface 170, and a wireless transceiver 172. The processing unit 110 may include one or more processing devices. In some embodiments, processing unit 110 may include an application processor 180, an image processor 190, or other suitable processing device. Similarly, the image capturing unit 120 may include any number of image capturing devices and components depending on the requirements of a particular application. In some embodiments, the image capture unit 120 may include one or more image capture devices (e.g., cameras) such as the image capture device 122, the image capture device 124, and the image capture device 126. The system 100 may also include a data interface 128 that communicatively connects the processing device 110 to the image capture device 120. For example, the data interface 128 may include one or more wired and / or wireless connections for transmitting the image data captured by the image training unit 120 to the processing unit 110.The wireless transceiver 172 may include one or more devices configured to exchange transmissions over an air interface to one or more networks (e.g., cellular, Internet, etc.) using a radio frequency, an infrared frequency, a magnetic field, or an electric field. The wireless transceiver 172 may use any known standard for transmitting and / or receiving data (e.g., Wi-Fi, Bluetooth® Bluetooth Smart, 802.15.4, ZigBee, etc.). Such transmissions may include communications from the host vehicle to one or more remote servers. Such transmissions may also include communications (one-way or two-way) between the host vehicle and one or more target vehicles in an environment of the host vehicle (e.g., to facilitate coordination of navigation of the host vehicle with respect to or together with target vehicles in the environment of the host vehicle), or even broadcast transmission to unspecified recipients in an environment of the transmitting vehicle.Each of the application processor 180 and the image processor 190 may include various types of processing devices. For example, one or both of the two application processors 180 and the image processor 190 may include a microprocessor, preprocessors (e.g., an image preprocessor), a graphics processing unit (GPU), a central processing unit (CPU), support circuits, digital signal processors, integrated circuits, memory, or other types of devices suitable for executing applications and image processing and analysis. In some embodiments, application processor 180 and / or image processor 190 may include any type of single-core or multi-core processor, microcontrollers for mobile devices, central processing unit, etc. Various processing devices may be used, such as processors from manufacturers such as Intel® AMD® etc., or GPUs from manufacturers such as NVIDIA® ATI® etc., which may include various architectures (e.g., x86 processor, ARM® etc.).In some embodiments, application processor 180 and / or image processor 190 may include any of the EyeQ series of processor chips available from Mobileye®. These processor designs each include multiple processing units with local memory and instruction sets. Such processors may include video inputs for receiving image data from multiple image sensors and may also include video output capabilities. In one example, the EyeQ2® uses 90 nm micron technology operating at 332 MHz. The EyeQ2® architecture consists of two floating point hyperthread 32-bit MIPS32® 34K® Cores (RISC CPUs), five vision computing engines (VCEs), three vector microcode processors (VMP®), Denali 64-bit Mobile DDR controllers, 128-bit internal sonics interconnect, dual 16-bit video input and 18-bit video output controllers, 16-channel DMA, and multiple peripheral devices. The MIPS34K CPU manages the five VCEs, three VMP™ and DMA, the second MIPS34K CPU and multi-channel DMA, as well as the other peripheral devices. The five VCEs, three VMP® and the MIPS34K CPU can perform intensive vision computations needed by multifunction cluster applications. In another example, the EyeQ3® which is a third generation processor and is six times more powerful than the EyeQ2® may be used in the disclosed embodiments. In other examples, the EyeQ4® and / or the EyeQ5® may be used in the disclosed embodiments. Of course, any newer or future EyeQ processing devices may also be used in conjunction with the disclosed embodiments.Any of the processing devices disclosed herein may be configured to perform certain functions. Configuring a processing device, such as any of the described EyeQ processors, or other controller or microprocessor, to perform certain functions may include programming computer-executable instructions and providing these instructions to the processing device for execution during operation of the processing device. In some embodiments, configuring a processing device may include programming the processing device directly with architectural instructions. For example, processing devices such as field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), and the like may be configured using one or more hardware description languages (HDLs).In other embodiments, configuring a processing device may include storing executable instructions in memory accessible by the processing device during operation. For example, the processing device may access the memory to obtain and execute the stored instructions during the operation. In any event, the processing device configured to perform the sensing, image analysis, and / or navigation functions disclosed herein represents a specialized hardware-based system that controls multiple hardware-based components of a host vehicle.Although FIG. 1 illustrates two separate processing devices included in processing unit 110, more or fewer processing devices may be used. For example, in some embodiments, a single processing device may be used to accomplish the tasks of application processor 180 and image processor 190. In other embodiments, these tasks may be performed by more than two processing devices. Further, in some embodiments, the system 100 may include one or more processing units 110 without including other components, such as the image capturing unit 120.The processing unit 110 may include various types of devices. For example, the processing unit 110 may include various devices such as a controller, an image preprocessor, a central processing unit (CPU), a graphics processing unit (GPU), support circuits, digital signal processors, integrated circuits, memory, or any other types of devices for image processing and analysis. The image preprocessor may include a video processor for capturing, digitizing, and processing the image material from the image sensors. The CPU may include any number of microcontrollers or microprocessors. The GPU may also include any number of microcontrollers or microprocessors. The support circuits may be any number of circuits generally well known in the art, including cache, power, clock and input-output circuits. The memory may store software that, when executed by the processor, controls operation of the system. The memory may include databases and image processing software. The memory may include any number of random access memories, read only memories, flash memories, disk drives, optical memory, tape memories, removable memories, and other types of memory. In one case, the memory may be separate from the processing unit 110. In another case, the memory may be integrated into the processing unit 110.Each memory 140, 150 may include software instructions that, when executed by a processor (e.g., application processor 180 and / or image processor 190), may control the operation of various aspects of the system 100. These storage units may include various databases and image processing software, as well as a trained system such as a neural network or a deep neural network. The storage units may include random access memory (RAM), read only memory (ROM), flash memory, disk drives, optical storage, tape storage, removable storage, and / or any other type of memory. In some embodiments, the storage units 140, 150 may be separate from the application processor 180 and / or image processor 190. In other embodiments, these memory units may be incorporated into the application processor 180 and / or the image processor 190.The position sensor 130 may include any type of device suitable for determining a location associated with at least one component of the system 100. In some embodiments, the position sensor 130 may include a GPS receiver. Such receivers can determine user position and velocity by processing signals broadcast from global positioning system satellites. Position information from the position sensor 130 may be provided to the application processor 180 and / or image processor 190.In some embodiments, the system 100 may include components such as a speed sensor (e.g., a tachometer, a speedometer) for measuring a speed of the vehicle 200, and / or an accelerometer (either single axis or multi-axis) for measuring the acceleration of the vehicle 200.The user interface 170 may include any device suitable for providing information to or receiving input from one or more users of the system 100. In some embodiments, user interface 170 may include user input devices including, for example, a touch screen, a microphone, a keyboard, pointing devices, road wheels, cameras, buttons, buttons, etc. With such input devices, a user may be able to provide information inputs or commands to system 100 by typing instructions or information, providing voice commands, selecting menu options on a screen using buttons, pointers, or eye tracking capabilities, or by any other suitable techniques for communicating information to system 100.The user interface 170 may be equipped with one or more processing devices configured to provide and receive information to or from a user and process that information for use by, for example, the application processor 180. In some embodiments, such processing devices may execute instructions to recognize and track eye movements, receive and interpret voice commands, recognize and interpret touches and / or gestures on a touch screen, respond to keyboard inputs or menu selections, etc. In some embodiments, the user interface 170 may include a display, a speaker, a tactile device, and / or any other devices for providing output information to a user.The map database 160 may include any type of database for storing map data useful for the system 100. In some embodiments, the map database 160 may include data relating to the location of various objects in a reference coordinate system, including roads, water features, geographic features, companies, points of interest, restaurant, gas stations, etc. Not only may the locations of such objects be stored in the map database 160 but descriptors relating to those objects, for example, names associated with one of the stored features. In some embodiments, the map database 160 may be physically located with other components of the system 100. Alternatively or additionally, the map database 160 or a portion thereof may be located remotely with respect to other components of the system 100 (e.g., processing unit 110). In such embodiments, information may be downloaded from the map database 160 over a wired or wireless data connection to a network (e.g., over a cellular network and / or the Internet, etc.). In some cases, the map database 160 may store a sparse data model that includes polynomial representations of certain road features (e.g., lane markings) or target trajectories for the host vehicle. Systems and methods for generating such a map are discussed below with reference to Figures 8-19.The image capture devices 122, 124, and 126 may each include any type of device suitable for capturing at least one image from an environment. Moreover, any number of image capture devices may be used to capture images for input to the image processor. Some embodiments may include only a single image capture device, while other embodiments may include two, three, or even four or more image capture devices. The image capture devices 122, 124, and 126 will be described in more detail below with reference to FIGS. 2B-2E.The system 100, or various components thereof, may be integrated into various different platforms. In some embodiments, the system 100 may be included in a vehicle 200, as shown in FIG. 2A. For example, the vehicle 200 may be equipped with a processing unit 110 and any of the other components of the system 100, as described above with respect to FIG. 1. While in some embodiments the vehicle 200 may be equipped with only a single image capture device (e.g., a camera), in other embodiments, multiple image capture devices may be used, such as those discussed in connection with FIGS. 2B-2E. For example, as shown in FIG. 2A, each of the image capture devices 122 and 124 of the vehicle 200 may be part of an Advanced Driver Assistance System (ADAS) imaging set.The image capturing devices included in the vehicle 200 as part of the image capturing unit 120 may be positioned at any suitable location. In some embodiments, as shown in FIGS. 2A-2E and 3A-3C, the image capture device 122 may be located near the rearview mirror. This position may provide a line of sight similar to that of the driver of the vehicle 200 that may aid in determining what is visible to the driver and not visible. The image capture device 122 may be positioned at any location near the rearview mirror, but placing the image capture device 122 on the driver side of the mirror may further aid in obtaining images representative of the driver's field of view and / or line of sight.Other locations for the image capture devices of the image capturing unit 120 may also be used. For example, the image capture device 124 may be located on or in a bumper of the vehicle 200. Such a location may be particularly suitable for image capture devices having a wide field of view. The line of sight of the bumper-mounted image capturing devices may be different from that of the driver, and thus the bumper image capturing device and the driver may not always see the same objects. The image capture devices (e.g., image capture devices 122, 124, and 126) may also be located in other locations. For example, the image capturing devices may be located on or in one or both of the side mirrors of the vehicle 200, the roof of the vehicle 200, the hood of the vehicle 200, the trunk of the vehicle 200, and the sides of the vehicle 200, mounted on one of the windows of the vehicle 200, positioned behind or in front of it, and mounted in or near light figures on the front and / or rear of the vehicle 200, etc.In addition to the image capture devices, the vehicle 200 may include various other components of the system 100. For example, the processing unit 110 may be included in the vehicle 200 that is either integrated with or separate from an engine control unit (ECU) of the vehicle. The vehicle 200 may also be equipped with a position sensor 130, such as a GPS receiver, and may also include a map database 160 and storage units 140 and 150.As discussed above, the wireless transceiver 172 may transmit and / or receive data over one or more networks (e.g., cellular networks, the Internet, etc.). For example, the wireless transceiver 172 may upload data collected by the system 100 to one or more servers and download data from the one or more servers. For example, via the wireless transceiver 172, the system 100 may receive periodic or as-needed updated data stored in the map database 160, the memory 140, and / or the memory 150. Similarly, the wireless transceiver 172 may upload any data (e.g., images captured by the image capture unit 120, data received from the position sensor 130, or other sensors, vehicle control systems, etc.) from the system 100 and / or any data processed by the processing unit 110 to the one or more servers.The system 100 may upload data to a server (e.g., the cloud) based on a privacy level setting. For example, the system 100 may implement privacy level settings to regulate or limit the types of data (including metadata) sent to the server that may uniquely identify a vehicle and / or a driver / owner of a vehicle. Such settings may be set by the user, for example, via the wireless transceiver 172, initialized by factory default settings, or by data received from the wireless transceiver 172.In some embodiments, the system 100 may upload data according to a "high" level of privacy, and upon setting, the system 100 may transmit data (e.g., location information regarding a route, captured images, etc.) without any details about the specific vehicle and / or the driver / owner. For example, when data is uploaded according to a "high" privacy setting, the system 100 may not include a chassis number (VIN) or a name of a driver or owner of the vehicle, and may instead transmit data such as captured images and / or limited location information related to a route.Other data protection levels are contemplated. For example, the system 100 may transmit data to a server according to a "medium" level of privacy and include additional information not included under a "high" level of privacy, such as a make and / or model of a vehicle and / or a type of vehicle (e.g., a passenger car, sport utility vehicle, truck, etc.). In some embodiments, system 100 may upload data according to a "low" level of data protection. Under a "low" privacy level setting, the system 100 may upload data and include information sufficient to uniquely identify a specific vehicle, owner / driver, and / or portion or all of a route being traveled by the vehicle. Such "low" privacy level data may include one or more of, for example, a VIN, a driver / owner name, a point of origin of a vehicle before departure, an intended destination of the vehicle, a make and / or model of the vehicle, a type of vehicle, etc.FIG. 2A is a schematic side view illustration of an example vehicle imaging system, in accordance with the disclosed embodiments. FIG. 2B is a schematic top view illustration of the embodiment shown in FIG. 2A. As illustrated in FIG. 2B, the disclosed embodiments may include a vehicle 200 that includes, in its body, a system 100 having a first image capture device 122 positioned proximate to the rearview mirror and / or proximate to the driver of the vehicle 200, a second image capture device 124 positioned on or within a bumper region (e.g., one of the bumper regions 210) of the vehicle 200, and a processing unit 110.As illustrated in FIG. 2C, the image capture devices 122 and 124 may both be positioned near the rearview mirror and / or near the driver of the vehicle 200. In addition, although two image capture devices 122 and 124 are shown in FIGS. 2B and 2C, it should be understood that other embodiments may include more than two image capture devices. For example, in the embodiments shown in FIGS. 2D and 2E, the first, second, and third image capture devices 122, 124, and 126 are included in the system 100 of the vehicle 200.As illustrated in FIG. 2D, the image capture device 122 may be positioned proximate to the rearview mirror and / or proximate to the driver of the vehicle 200, and the image capture devices 124 and 126 may be positioned on or within a bumper region (e.g., one of the bumper regions 210) of the vehicle 200. And as shown in FIG. 2E, the image capture devices 122, 124, and 126 may be positioned proximate the rearview mirror and / or proximate the driver's seat of the vehicle 200. The disclosed embodiments are not limited to any particular number and configuration of image capture devices, and the image capture devices may be positioned at any suitable location within and / or on the vehicle 200.It should be understood that the disclosed embodiments are not limited to vehicles and could be applied in other contexts. It should also be understood that the disclosed embodiments are not limited to a particular type of vehicle 200 and may be applicable to all types of vehicles, including automobiles, trucks, trailers, and other types of vehicles.The first image capture device 122 may include any suitable type of image capture device. The image capture device 122 may include an optical axis. In one case, the image capture device 122 may include an Aptina M9V024 global shutter type WVGA sensor. In other embodiments, the image capture device 122 may provide a resolution of 1280x960 pixels and include a rolling shutter. The image capturing device 122 may include various optical elements. In some embodiments, one or more lenses may be included to provide, for example, a desired focal length and field of view to the image capture device. In some embodiments, the image capture device 122 may be associated with a 6 mm lens or a 12 mm lens. In some embodiments, the image capture device 122 may be configured to capture images having a desired field of view (FOV) 202, as illustrated in FIG. 2D. For example, the image capture device 122 may be configured to have a regular FOV, such as within a range of 40 degrees to 56 degrees, including a 46 degree FOV, 50 degree FOV, 52 degree FOV, or greater. Alternatively, the image capture device 122 may be configured to have a narrow FOV in the range of 23 to 40 degrees, such as a 28-degree FOV or 36-degree FOV. In addition, the image capturing device 122 may be configured to have a wide FOV in the range of 100 to 180 degrees. In some embodiments, the image capture device 122 may include a wide-angle bumper camera or one with up to a 180-degree FOV. In some embodiments, the image capture device 122 may be a 7.2 M pixel image capture device with an aspect ratio of about 2:1 (e.g., HxV=3800 x 1900 pixels) with about 100 degrees horizontal FOV. Such an image capturing device may be used instead of a configuration having three image capturing devices. Due to significant lens distortion, the vertical FOV of such an image capture device may be significantly less than 50 degrees in implementations where the image capture device uses a radially symmetric lens. For example, such a lens may not be radially symmetric, which would allow a vertical FOV of more than 50 degrees with a horizontal FOV of 100 degrees.The first image capture device 122 may capture a plurality of first images relative to a scene associated with the vehicle 200. Each of the plurality of first images may be acquired as a series of image scan lines that may be captured using a rolling shutter. Each scan line may include a plurality of pixels.The first image capture device 122 may have a sampling rate associated with the capture of each of the first series of image scanlines. The sampling rate may refer to a rate at which an image sensor may capture image data associated with each pixel included in a particular scan line.The image capture devices 122, 124, and 126 may include any suitable type and number of image sensors, including, for example, CCD sensors or CMOS sensors. In one embodiment, a CMOS image sensor may be employed along with a rolling shutter such that each pixel in a row is read sequentially and scanning of the rows is performed on a row-by-row basis until an entire frame is acquired. In some embodiments, the rows may be sequentially detected from top to bottom with respect to the frame.In some embodiments, one or more of the image capture devices disclosed herein (e.g., image capture devices 122, 124, and 126) may form a high resolution imager and have a resolution of greater than 5 M pixels, 7 M pixels, 10 M pixels, or more.The use of a rolling shutter may result in pixels in different rows being exposed and captured at different times, which may result in distortion and other image artifacts in the captured frame. On the other hand, if the image capture device 122 is configured to operate with global or synchronous shutter, all pixels may be exposed for the same time and during a common exposure period. As a result, the image data in a frame collected by a system using global shutter represents a snapshot of the entire FOV (such as FOV 202) at a particular time. In contrast, in a rolling shutter application, each row in a single image is exposed and data is acquired at different times. Therefore, moving objects may appear distorted in an image capturing device with a rolling shutter. This phenomenon will be described in more detail below.The second image capture device 124 and the third image capture device 126 may be any type of image capture device. Like the first image capture device 122, each of the image capture devices 124 and 126 may include an optical axis. In one embodiment, each of the image capture devices 124 and 126 may include an Aptina M9V024 global shutter type WVGA sensor. Alternatively, each of the image capturing devices 124 and 126 may include a rolling shutter. Like the image capture device 122, the image capture devices 124 and 126 may be configured to include various lenses and optical elements. In some embodiments, the lenses associated with the image capture devices 124 and 126 may provide FOV (such as FOVs 204 and 206) that are equal to or narrower than a FOV (such as FOV 202) associated with the image capture device 122. For example, the image capture devices 124 and 126 may have FOVs of 40 degrees, 30 degrees, 26 degrees, 23 degrees, 20 degrees, or less.The image capture devices 124 and 126 may capture a plurality of second and third images relative to a scene associated with the vehicle 200. Each of the plurality of second and third images may be captured as second and third series of image scan lines that may be captured using a rolling shutter. Each scan line or row may include a plurality of pixels. The image capture devices 124 and 126 may have second and third sampling rates associated with the capture of each of the image scanlines included in the second and third rows.Each image capture device 122, 124, and 126 may be positioned at any suitable position and orientation relative to the vehicle 200. The relative positioning of the image capture devices 122, 124, and 126 may be selected to aid in merging the information captured by the image capture devices. For example, in some embodiments, a FOV (such as FOV 204) associated with image capture device 124 may partially or fully overlap a FOV (such as FOV 202) associated with image capture device 122 and a FOV (such as FOV 206) associated with image capture device 126.The image capture devices 122, 124, and 126 may be located on the vehicle 200 at any suitable relative heights. In one case, there may be a height difference between the image capturing devices 122, 124, and 126 that may provide sufficient parallax information to enable stereo analysis. For example, as shown in FIG. 2A, the two image capturing devices 122 and 124 are at different heights. There may also be a lateral displacement difference between the image capture devices 122, 124, and 126, which provides additional parallax information for stereo analysis by the processing unit 110, for example. The difference in lateral displacement may be denoted by d x as shown in FIGS. 2C and 2D. In some embodiments, a forward or backward shift (e.g., range shift) may exist between the image capture devices 122, 124, and 126. For example, the image capture device 122 may be located 0.5 to 2 meters or more behind the image capture device 124 and / or the image capture device 126. This type of displacement may allow one of the image capture devices to cover potential blind spots of the other image capture device(s).The image capture devices 122 may have any suitable resolution capability (e.g., number of pixels associated with the image sensor), and the resolution of the image sensor(s) associated with the image capture device 122 may be higher, lower, or equal to the resolution of the image sensor(s) associated with the image capture devices 124 and 126. In some embodiments, the image sensor(s) associated with the image capture device 122 and / or the image capture devices 124 and 126 may have a resolution of 640 x 480, 1024 x 768, 1280 x 960, or any other suitable resolution.The frame rate (e.g., the rate at which an image capture device captures a set of pixel data of a frame before moving to capture pixel data associated with the next frame) may be controllable. The frame rate associated with the image capture device 122 may be higher, lower, or equal to the frame rate associated with the image capture devices 124 and 126. The frame rate associated with the image capture devices 122, 124, and 126 may depend on a variety of factors that may affect the timing of the frame rate. For example, one or more of the image capture devices 122, 124, and 126 may include a selectable pixel delay time period imposed before or after the capture of image data associated with one or more pixels of an image sensor in the image capture device 122, 124, and / or 126. Generally, image data corresponding to each pixel may be captured according to a clock rate for the device (e.g., one pixel per clock cycle). Additionally, in embodiments including a rolling shutter, one or more of the image capture devices 122, 124, and 126 may include a selectable horizontal blanking period imposed before or after capturing image data associated with a row of pixels of an image sensor in the image capture device 122, 124, and / or 126. Further, one or more of the image capture devices 122, 124, and / or 126 may include a selectable vertical blanking period imposed before or after capturing image data associated with a frame of the image capture device 122, 124, and 126.These timings may enable synchronization of frame rates associated with the image capture devices 122, 124, and 126, even if the line scan rates are different, respectively. Additionally, as will be discussed in more detail below, these selectable timings may enable synchronization of image capture from a region where the FOV of image capture device 122 overlaps one or more FOVs of image capture devices 124 and 126, even though the field of view of image capture device 122 is different than the FOVs of image capture devices 124 and 126, among other factors (e.g., image sensor resolution, maximum line scan rates, etc.).The timing of the frame rate in the image capture device 122, 124, and 126 may depend on the resolution of the associated image sensors. For example, if similar line scan rates are assumed for both devices, if one device includes an image sensor with a resolution of 640 x 480 and another device includes an image sensor with a resolution of 1280 x 960, then more time is required to capture a single image of image data from the sensor with the higher resolution.Another factor that may affect the timing of image data acquisition in the image acquisition devices 122, 124, and 126 is the maximum line scan rate. For example, capturing a series of image data from an image sensor included in the image capturing device 122, 124, and 126 requires a minimum amount of time. Assuming no pixel delay periods are added, this minimum time duration for capturing a series of image data refers to the maximum line scan rate for a particular device. Devices offering higher maximum line scan rates have the potential to provide higher frame rates than devices with lower maximum line scan rates. In some embodiments, one or more of the image capture devices 124 and 126 may have a maximum line scan rate that is higher than a maximum line scan rate associated with the image capture device 122. In some embodiments, the maximum line scan rate of the image capture device 124 and / or 126 may be 1.25, 1.5, 1.75, or 2 times or more a maximum line scan rate of the image capture device 122.In another embodiment, the image capture devices 122, 124, and 126 may have the same maximum line sample rate, but the image capture device 122 may operate at a sample rate that is less than or equal to its maximum sample rate. The system may be configured such that one or more of the image capture devices 124 and 126 operate at a line scan rate equal to the line scan rate of the image capture device 122. In other cases, the system may be configured such that the line scan rate of the image capture device 124 and / or the image capture device 126 may be 1.25, 1.5, 1.75, or 2 times or more the line scan rate of the image capture device 122.In some embodiments, the image capture devices 122, 124, and 126 may be asymmetric. That is, they may include cameras having different fields of view (FOV) and focal lengths. The fields of view of the image capture devices 122, 124, and 126 may include, for example, any desired range with respect to an environment of the vehicle 200. In some embodiments, one or more of the image capture devices 122, 124, and 126 may be configured to capture image data from an environment in front of the vehicle 200, behind the vehicle 200, at the sides of the vehicle 200, or combinations thereof.Further, the focal length associated with each image capture device 122, 124, and / or 126 may be selectable (e.g., by including appropriate lenses, etc.) such that each device captures images of objects in a desired range of distance with respect to the vehicle 200. For example, in some embodiments, the image capture devices 122, 124, and 126 may capture images of close-up objects within a few meters of the vehicle. The image capture devices 122, 124, and 126 may also be configured to capture images of objects at distances farther from the vehicle (e.g., 25 m, 50 m, 100 m, 150 m, or more). Further, the focal lengths of the image capture devices 122, 124, and 126 may be selected such that one image capture device (e.g., image capture device 122) may capture images of objects relatively close to the vehicle (e.g., within 10 m or within 20 m), while the other image capture devices (e.g., image capture devices 124 and 126) may capture images of objects farther away (e.g., greater than 20 m, 50 m, 100 m, 150 m, etc.) from the vehicle 200.According to some embodiments, the FOV of one or more image capture devices 122, 124, and 126 may have a wide angle. For example, it may be advantageous to have a FOV of 140 degrees, particularly for the image capture devices 122, 124, and 126, which may be used to capture images of the area proximate the vehicle 200. For example, the image capture device 122 may be used to capture images of the area to the right or left of the vehicle 200, and in such embodiments, it may be desirable for the image capture device 122 to have a wide FOV (e.g., at least 140 degrees).The field of view associated with each of the image capture devices 122, 124, and 126 may depend on the respective focal lengths. For example, as the focal length increases, the corresponding field of view decreases.The image capture devices 122, 124, and 126 may be configured to have any suitable fields of view. In a particular example, the image capture device 122 may have a horizontal FOV of 46 degrees, the image capture device 124 may have a horizontal FOV of 23 degrees, and the image capture device 126 may have a horizontal FOV between 23 and 46 degrees. In another case, the image capture device 122 may have a horizontal FOV of 52 degrees, the image capture device 124 may have a horizontal FOV of 26 degrees, and the image capture device 126 may have a horizontal FOV between 26 and 52 degrees. In some embodiments, a ratio of the FOV of the image capture device 122 to the FOVs of the image capture device 124 and / or the image capture device 126 may vary from 1.5 to 2.0. In other embodiments, this ratio may vary between 1.25 and 2.25.The system 100 may be configured such that a field of view of the image capture device 122 overlaps, at least in part or completely, with a field of view of the image capture device 124 and / or the image capture device 126. In some embodiments, system 100 may be configured such that the fields of view of image capture devices 124 and 126 fall within (e.g., are narrower than) and share a common center with, the field of view of image capture device 122, for example. In other embodiments, the image capture devices 122, 124, and 126 may capture adjacent FOVs or have partial overlap in their FOVs. In some embodiments, the fields of view of the image capture devices 122, 124, and 126 may be oriented such that a center of the narrower field of view image capture devices 124 and / or 126 may be located in a lower half of the wider field of view image capture device 122.FIG. 2F is a schematic illustration of example vehicle control systems, in accordance with the disclosed embodiments. As indicated in FIG. 2F, the vehicle 200 may include a throttle system 220, a brake system 230, and a steering system 240. The system 100 may provide inputs (e.g., control signals) to one or more of a throttle system 220, a brake system 230, and a steering system 240 via one or more data connections (e.g., any wired and / or wireless connection or connections for transmitting data). For example, based on an analysis of images captured by the image capture devices 122, 124, and / or 126, the system 100 may provide control signals to one or more of a throttle system 220, a brake system 230, and a steering system 240 to navigate the vehicle 200 (e.g., by causing acceleration, turn, lane shift, etc.). Further, the system 100 may receive inputs from one or more of a throttle system 220, a brake system 230, and a steering system 24 indicative of operating conditions of the vehicle 200 (e.g., speed, whether the vehicle 200 is braking and / or turning, etc.). Further details are provided in connection with the following FIGS. 4-7.As shown in FIG. 3A, the vehicle 200 may also include a user interface 170 for interacting with a driver or a passenger of the vehicle 200. For example, in a vehicle application, the user interface 170 may include a touch screen 320, buttons 330, buttons 340, and a microphone 350. A driver or passenger of the vehicle 200 may also use handles (e.g., located on or near the steering column of the vehicle 200, including, for example, turn signal handles), buttons (e.g., located on the steering wheel of the vehicle 200), and the like to interact with the system 100. In some embodiments, the microphone 350 may be positioned adjacent to a rearview mirror 310. Similarly, in some embodiments, the image capture device 122 may be located near the rearview mirror 310. In some embodiments, the user interface 170 may also include one or more speakers 360 (e.g., speakers of a vehicle audio system). For example, the system 100 may provide various notifications (e.g., alerts) via the speakers 360.FIGS. 3B-3D are illustrations of an example camera mount 370 configured to be positioned behind a rearview mirror (e.g., rearview mirror 310) and against a vehicle windshield, in accordance with the disclosed embodiments. As shown in FIG. 3B, the camera mount 370 may include image capture devices 122, 124, and 126. The image capture devices 124 and 126 may be positioned behind an visor 380, which may be flush with the vehicle windshield and may include a composition of film and / or anti-reflective materials. For example, the visor 380 may be positioned such that the visor aligns with a vehicle windshield having a matched slope. In some embodiments, each of the image capture devices 122, 124, and 126 may be positioned behind the visor 380, as shown in FIG. 3D, for example. The disclosed embodiments are not limited to any particular configuration of the image capture devices 122, 124, and 126, the camera mount 370, and the visor 380. FIG. 3C is an illustration of the camera mount 370 shown in FIG. 3B from a front perspective.As will be apparent to those skilled in the art having the benefit of this disclosure, numerous variations and / or modifications may be made to the embodiments disclosed above. For example, not all components are essential to the operation of the system 100. Further, any component may be located in any suitable portion of the system 100, and the components may be rearranged into a variety of configurations while providing the functionality of the disclosed embodiments. Thus, the above configurations are examples and, regardless of the configurations discussed above, the system 100 may provide a wide range of functionality to analyze the environment of the vehicle 200 and navigate the vehicle 200 in response to the analysis.As discussed in greater detail below and in accordance with various disclosed embodiments, the system 100 may provide a variety of features related to autonomous driving and / or driver assistance technology. For example, the system 100 may analyze image data, position data (e.g., GPS location information), map data, speed data, and / or data from sensors included in the vehicle 200. The system 100 may collect the data for analysis from, for example, the image capturing unit 120, the position sensor 130, and other sensors. Further, the system 100 may analyze the collected data to determine whether or not the vehicle 200 should perform a particular action, and then automatically perform the particular action without human intervention. For example, if the vehicle 200 navigates without human intervention, the system 100 may automatically control braking, accelerating, and / or steering of the vehicle 200 (e.g., by sending control signals to one or more of a throttle system 220, a brake system 230, and a steering system 240). Further, the system 100 may analyze the collected data and issue warnings and / or alarms to vehicle occupants based on the analysis of the collected data. Additional details regarding the various embodiments provided by the system 100 are provided below.Forward Multi-Image SystemAs discussed above, the system 100 may provide driving assistance functionality using a multi-camera system. The multi-camera system may use one or more cameras directed in the forward direction of a vehicle. In other embodiments, the multi-camera system may include one or more cameras facing the side of a vehicle or the rear of the vehicle. For example, in one embodiment, system 100 may use a two-camera imaging system, where a first camera and a second camera (e.g., image capture devices 122 and 124) may be positioned at the front and / or sides of a vehicle (e.g., vehicle 200). The first camera may have a field of view that is greater than, less than, or partially overlapping the field of view of the second camera. In addition, the first camera may be connected to a first image processor to perform monocular image analysis of images provided by the first camera, and the second camera may be connected to a second image processor to perform monocular image analysis of images provided by the second camera. The outputs (e.g., processed information) of the first and second image processors may be combined. In some embodiments, the second image processor may receive images from both the first camera and the second camera to perform stereo analysis. In another embodiment, the system 100 may use a three camera imaging system, each of the cameras having a different field of view. Such a system can therefore make decisions based on information derived from objects located at varying distances both forward and to the sides of the vehicle. References to monocular image analysis may refer to cases where image analysis is performed based on images captured from a single viewpoint (e.g., from a single camera). Stereo image analysis may refer to cases where image analysis is performed based on two or more images acquired with one or more variations of an image acquisition parameter. For example, captured images suitable for performing stereo image analysis may include images captured from two or more different positions, from different fields of view, using different focal lengths, along with parallax information, etc.For example, in one embodiment, the system 100 may implement a three camera configuration using the image capture devices 122, 124, and 126. In such a configuration, the image capture device 122 may provide a narrow field of view (e.g., 34 degrees or other values selected from a range of about 20 to 45 degrees, etc.), the image capture device 124 may provide a wide field of view (e.g., 150 degrees or other values selected from a range of about 100 to about 180 degrees), and the image capture device 126 may provide an intermediate field of view (e.g., 46 degrees or other values selected from a range of about 35 to about 60 degrees). In some embodiments, the image capture device 126 may function as a master or primary camera. The image capture devices 122, 124, and 126 may be positioned behind the rearview mirror 310 and positioned substantially side-by-side (e.g., 6 cm apart). Further, in some embodiments, as discussed above, one or more of the image capture devices 122, 124, and 126 may be mounted behind the visor 380, which is flush with the windshield of the vehicle 200. Such shielding may serve to minimize the impact of reflections from inside the car on the image capture devices 122, 124, and 126.In another embodiment, as discussed above in connection with FIGS. 3B and 3C, the wide field of view camera (e.g., image capture device 124 in the above example) may be mounted lower than the narrow and main field of view cameras (e.g., image capture devices 122 and 126 in the above example). This configuration may provide a clear line of sight from the wide field of view camera. To reduce reflections, the cameras may be mounted near the windshield of the vehicle 200 and include polarizers on the cameras to attenuate reflected light.A system with three cameras may provide certain performance characteristics. For example, some embodiments may include a capability to validate the detection of objects by one camera based on detection results from another camera. For example, in the three camera configuration discussed above, the processing unit 110 may include three processing devices (e.g., three EyeQ rows of processor chips, as discussed above), each processing device dedicated to processing images captured by one or more of the image capture devices 122, 124, and 126.In a system with three cameras, a first processing device may receive images from both the main camera and the narrow field of view camera and perform image processing of the narrow FOV camera to detect, for example, other vehicles, pedestrians, lane markings, traffic signs, traffic lights, and other road objects. Further, the first processing device may calculate a disparity of pixels between the images from the main camera and the narrow camera and generate a 3D reconstruction of the environment of the vehicle 200. The first processing device may then combine the 3D reconstruction with 3D map data or with 3D information calculated based on information from another camera.The second processing device may receive images from the master camera and perform image processing to detect other vehicles, pedestrians, lane markings, traffic signs, traffic lights, and other road objects. Additionally, the second processing device may calculate a camera displacement and calculate a disparity of pixels between successive images based on the displacement and generate a 3D reconstruction of the scene (e.g., a texture of motion). The second processing device may send the structure from a motion-based 3D reconstruction to the first processing device to be combined with the stereo 3D images.The third processing device may receive images from the wide FOV camera and process the images to detect vehicles, pedestrians, lane markings, traffic signs, traffic lights, and other road objects. The third processing device may further execute additional processing instructions to analyze images to identify objects moving in the image, such as vehicles changing lanes, pedestrians, etc.In some embodiments, streams of image-based information that are independently captured and processed may provide a way to provide redundancy in the system. Such redundancy may include, for example, the use of a first image capturing device and the images processed by this device to validate and / or supplement information obtained by capturing and processing image information from at least one second image capturing device.In some embodiments, system 100 may use two image capture devices (e.g., image capture devices 122 and 124) to provide navigation assistance to vehicle 200 and use a third image capture device (e.g., image capture device 126) to provide redundancy and validate the analysis of data received from the other two image capture devices. For example, in such a configuration, image capture devices 122 and 124 may provide images for stereo analysis by system 100 for navigating vehicle 200, while image capture device 126 may provide images for monocular analysis by system 100 to provide redundancy and validation of information obtained based on images captured by image capture device 122 and / or image capture device 124. That is, the image capture device 126 (and a corresponding processing device) may be considered a redundant subsystem for providing a check on the analysis derived from the image capture devices 122 and 124 (e.g., to provide an automatic emergency braking (AEB) system). Moreover, in some embodiments, redundancy and validation of the received data may be supplemented based on information received from one or more sensors (e.g., radar, lidar, acoustic sensors, information received from one or more transceivers external to a vehicle, etc.).One skilled in the art will recognize that the above camera configurations, camera placements, number of cameras, camera locations, etc. are only examples. These components and others described with respect to the overall system may be assembled and used in a variety of different configurations without departing from the scope of the disclosed embodiments. Further details regarding the use of a multi-camera system to provide driver assistance and / or autonomous vehicle functionality follow below.FIG. 4 is an example functional block diagram of memory 140 and / or 150 that may be stored / programmed with instructions to perform one or more operations, in accordance with the disclosed embodiments. Although the following refers to the memory 140, one skilled in the art will recognize that instructions may be stored in the memory 140 and / or 150.As shown in FIG. 4, the memory 140 may store a monocular image analysis module 402, a stereo image analysis module 404, a speed and acceleration module 406, and a navigation response module 408. The disclosed embodiments are not limited to any particular configuration of the memory 140. Further, the application processor 180 and / or the image processor 190 may execute the instructions stored in any of the modules 402, 404, 406, and 408 included in the memory 140. One skilled in the art will understand that references in the discussions that follow to processing unit 110 may refer individually or collectively to application processor 180 and image processor 190. Accordingly, steps of any of the following processes may be performed by one or more processing devices.In one embodiment, monocular image analysis module 402 may store instructions (such as computer vision software) that, when executed by processing unit 110, perform monocular image analysis of a set of images captured by one of image capture devices 122, 124, and 126. In some embodiments, the processing unit 110 may combine information from a set of images with additional sensory information (e.g., radar, lidar, etc.) to perform monocular image analysis. As described in connection with subsequent FIGS. 5A-5D, monocular image analysis module 402 may include instructions for detecting a set of features within the set of images, such as lane markings, vehicles, pedestrians, road signs, highway exit ramps, traffic lights, hazardous objects, and any other feature associated with a vehicle's environment. Based on the analysis, the system 100 may cause (e.g., via the processing unit 110) one or more navigation responses in the vehicle 200, such as a turn, a lane shift, a change in acceleration, and the like, as discussed below in connection with the navigation response module 408.In one embodiment, stereo image analysis module 404 may store instructions (such as computer vision software) that, when executed by processing unit 110, perform stereo image analysis of first and second sets of images captured by a combination of image capture devices selected from any of image capture devices 122, 124, and 126. In some embodiments, the processing unit 110 may combine information from the first and second sets of images with additional sensory information (e.g., information from radar) to perform the stereo image analysis. For example, the stereo image analysis module 404 may include instructions to perform stereo image analysis based on a first set of images captured by the image capture device 124 and a second set of images captured by the image capture device 126. As described in connection with the following FIG. 6, the stereo image analysis module 404 may include instructions to detect a set of features within the first and second sets of images, such as lane markings, vehicles, pedestrians, road signs, highway exit ramps, traffic lights, hazardous objects, and the like. Based on the analysis, the processing unit 110 may cause one or more navigation responses in the vehicle 200, such as a turn, a lane shift, a change in acceleration, and the like, as discussed below in connection with the navigation response module 408. Further, in some embodiments, stereo image analysis module 404 may implement techniques associated with a trained system (such as a neural network or a deep neural network) or an untrained system, such as a system that may be configured to use computer vision algorithms to recognize and / or designate objects in an environment from which sensory information was captured and processed. In one embodiment, stereo image analysis module 404 and / or other image processing modules may be configured to use a combination of a trained and non-trained system.In one embodiment, the speed and acceleration module 406 may store software configured to analyze data received from one or more computing devices and electromechanical devices in the vehicle 200 configured to cause a change in speed and / or acceleration of the vehicle 200. For example, the processing unit 110 may execute instructions associated with the speed and acceleration module 406 to calculate a target speed for the vehicle 200 based on data derived from execution of the monocular image analysis module 402 and / or the stereo image analysis module 404. Such data may include, for example, a target position, speed, and / or acceleration, the position and / or speed of the vehicle 200 with respect to a nearby vehicle, a pedestrian or a road object, position information for the vehicle 200 with respect to lane markings of the road, and the like. Additionally, the processing unit 110 may calculate a target speed for the vehicle 200 based on sensory inputs (e.g., information from radar) and inputs from other systems of the vehicle 200, such as the throttle system 220, the brake system 230, and / or the steering system 240 of the vehicle 200. Based on the calculated target speed, the processing unit 110 may transmit electronic signals to the throttle system 220, the brake system 230, and / or the steering system 240 of the vehicle 200 to trigger a change in speed and / or acceleration by, for example, physically depressing the brake or releasing the accelerator of the vehicle 200.In one embodiment, the navigation response module 408 may store software executable by the processing unit 110 to determine a desired navigation response based on data derived from execution of the monocular image analysis module 402 and / or the stereo image analysis module 404. Such data may include position and speed information associated with nearby vehicles, pedestrians, and road objects, target position information for the vehicle 200, and the like. Additionally, in some embodiments, the navigation response may be (partially or fully) based on map data, a predetermined position of the vehicle 200, and / or a relative speed or acceleration between the vehicle 200 and one or more objects detected by execution of the monocular image analysis module 402 and / or the stereo image analysis module 404. The navigation response module 408 may also determine a desired navigation response based on sensory inputs (e.g., radar information) and inputs from other systems of the vehicle 200, such as the throttle system 220, the brake system 230, and the steering system 240 of the vehicle 200. Based on the desired navigation response, the processing unit 110 may transmit electronic signals to the throttle system 220, the brake system 230, and the steering system 240 of the vehicle 200 to trigger a desired navigation response, for example, by steering the steering wheel of the vehicle 200 to achieve a rotation of a predetermined angle. In some embodiments, the processing unit 110 may use the output of the navigation response module 408 (e.g., the desired navigation response) as an input to execute the speed and acceleration module 406 to calculate a change in speed of the vehicle 200.Moreover, each of the modules disclosed herein (e.g., modules 402, 404, and 406) may implement techniques associated with a trained system (such as a neural network or a deep neural network) or an untrained system.FIG. 5A is a flow diagram illustrating an example process 500A for causing one or more navigation responses based on monocular image analysis, in accordance with the disclosed embodiments. At step 510, the processing unit 110 may receive a plurality of images via the data interface 128 between the processing unit 110 and the image capturing unit 120. For example, a camera included in the image capturing unit 120 (such as the image capturing device 122 having the field of view 202) may capture a plurality of images of an area in front of the vehicle 200 (or, for example, to the sides or the rear of a vehicle) and transmit them to the processing unit 110 via a data connection (e.g., digital, wired, USB, wireless, Bluetooth, etc.). The processing unit 110 may execute the monocular image analysis module 402 to analyze the plurality of images at step 520, as described in more detail below in connection with FIGS. 5B-5D. By performing the analysis, the processing unit 110 may recognize a set of features within the set of images, such as lane markings, vehicles, pedestrians, road signs, highway exit ramps, traffic lights, and the like.The processing unit 110 may also execute the monocular image analysis module 402 to detect various road hazards at step 520, such as, for example, portions of a truck tire, fallen road signs, loose cargo, small animals, and the like. Road hazards may vary in structure, shape, size, and color, which may make detection of such hazards more difficult. In some embodiments, the processing unit 110 may execute the monocular image analysis module 402 to perform multiframe analysis on the plurality of images to detect road hazards. For example, the processing unit 110 may estimate the camera motion between successive frames and calculate the disparity in pixels between the frames to generate a 3D map of the road. The processing unit 110 may then use the 3D map to detect the road surface as well as the hazards existing over the road surface.At step 530, the processing unit 110 may execute the navigation response module 408 to generate one or more navigation responses in the vehicle 200 based on the analysis performed at step 520 and the techniques described above in connection with FIG. 4. Navigation responses may include, for example, a turn, a lane shift, a change in acceleration, and the like. In some embodiments, the processing unit 110 may use data derived from execution of the speed and acceleration module 406 to cause the one or more navigation responses. In addition, multiple navigation reactions may occur simultaneously, sequentially, or in any combination thereof. For example, the processing unit 110 may cause the vehicle 200 to switch a lane and then accelerate by, for example, sequentially transmitting control signals to the steering system 240 and the throttle system 220 of the vehicle 200. Alternatively, the processing unit 110 may cause the vehicle 200 to brake while simultaneously switching lanes, for example, by simultaneously transmitting control signals to the brake system 230 and the steering system 240 of the vehicle 200.FIG. 5B is a flowchart illustrating an example process 500B for detecting one or more vehicles and / or pedestrians in a set of images, in accordance with the disclosed embodiments. The processing unit 110 may execute the monocular image analysis module 402 to implement the process 500B. At step 540, the processing unit 110 may determine a set of candidate objects representing potential vehicles and / or pedestrians. For example, the processing unit 110 may scan one or more images, compare the images to one or more predetermined patterns, and identify within each image possible locations that may include objects of interest (e.g., vehicles, pedestrians, or portions thereof). The predetermined patterns may be configured to achieve a high rate of "false hits" and a low rate of "misses". For example, the processing unit 110 may use a low threshold of similarity with predetermined patterns to identify candidate objects as possible vehicles or pedestrians. This may allow the processing unit 110 to reduce the likelihood that a candidate object representing a vehicle or pedestrian is missing (e.g., not identified).At step 542, the processing unit 110 may filter the set of candidate objects to exclude certain candidates (e.g., irrelevant or less relevant objects) based on classification criteria. Such criteria may be derived from various characteristics associated with object types stored in a database (e.g., a database stored in memory 140). Characteristics may include object shape, dimensions, texture, position (e.g., relative to the vehicle 200), and the like. Thus, the processing unit 110 may use one or more sets of criteria to reject incorrect candidates from the set of candidate objects.At step 544, the processing unit 110 may analyze multiple frames of images to determine whether objects in the set of candidate objects represent vehicles and / or pedestrians. For example, the processing unit 110 may track a detected candidate object over successive frames and accumulate frame data associated with the detected object (e.g., size, position relative to the vehicle 200, etc.). In addition, the processing unit 110 may estimate parameters for the detected object and compare the frame position data of the object with a predicted position.At step 546, the processing unit 110 may generate a set of measurements for the detected objects. Such measurements may include, for example, position, velocity, and acceleration values (relative to the vehicle 200) associated with the detected objects. In some embodiments, the processing unit 110 may generate the measurements based on estimation techniques using a series of time-based observations, such as Kalman filters or linear quadratic estimation (LQE), and / or based on available modeling data for different object types (e.g., cars, trucks, pedestrians, bicycles, road signs, etc.). The Kalman filters may be based on a measurement of the scale of an object, where the scale measurement is proportional to a time to collision (e.g., the amount of time the vehicle 200 takes to reach the object). Thus, by performing steps 540- 546, the processing unit 110 may identify vehicles and pedestrians appearing within the set of captured images and derive information (e.g., position, speed, size) associated with the vehicles and pedestrians. Based on the identification and the derived information, the processing unit 110 may cause one or more navigation responses in the vehicle 200 as described above in connection with FIG. 5A.At step 548, the processing unit 110 may perform optical flow analysis of one or more images to reduce the probabilities that a "false hit" is detected and a candidate object representing a vehicle or pedestrian is missing. The optical flow analysis may relate to, for example, analyzing motion patterns relative to the vehicle 200 in one or more images associated with other vehicles and pedestrians that are different from road surface motion. The processing unit 110 may calculate the movement of candidate objects by observing the different positions of the objects over a plurality of frames acquired at different times. The processing unit 110 may use the position and time values as inputs to mathematical models to calculate the motion of the candidate objects. Thus, the optical flow analysis may provide another method for detecting vehicles and pedestrians located near the vehicle 200. The processing unit 110 may perform the optical flow analysis in combination with steps 540- 546 to provide redundancy for detecting vehicles and pedestrians and to increase the reliability of the system 100.FIG. 5C is a flow diagram illustrating an example process 500C for detecting road markings and / or lane geometry information in a set of images, in accordance with the disclosed embodiments. The processing unit 110 may execute the monocular image analysis module 402 to implement the process 500C. At step 550, the processing unit 110 may detect a set of objects by scanning one or more images. To detect segments of lane markings, lane geometry information, and other relevant road markings, the processing unit 110 may filter the set of objects to exclude those determined to be irrelevant (e.g., small holes, small rocks, etc.). At step 552, the processing unit 110 may group the segments that belong to the same road marking or lane marking detected at step 550. Based on the grouping, the processing unit 110 may develop a model to represent the detected segments, such as a mathematical model.At step 554, the processing unit 110 may construct a set of measurements associated with the detected segments. In some embodiments, the processing unit 110 may generate a projection of the detected segments from the image plane onto the real plane. The projection may be characterized using a third degree polynomial having coefficients corresponding to physical properties such as the position, slope, curvature, and curvature derivative of the detected road. In generating the projection, the processing unit 110 may take into account changes in the road surface as well as pitch and roll rates associated with the vehicle 200. In addition, the processing unit 110 may model the road height by analyzing position and motion cues present on the road surface. Further, the processing unit 110 may estimate the pitch and roll rates associated with the vehicle 200 by tracking a set of feature points in the one or more images.At step 556, processing unit 110 may perform multiframe analysis by, for example, tracking the detected segments over successive frames and accumulating frame data associated with detected segments. Because the processing unit 110 performs multiframe analysis, the set of measurements made at step 554 may become more reliable and associated with an ever-higher confidence level. Thus, by performing steps 550, 552, 554, and 556, processing unit 110 may identify road markings that appear within the set of captured images and derive lane geometry information. Based on the identification and the derived information, the processing unit 110 may cause one or more navigation responses in the vehicle 200 as described above in connection with FIG. 5A.At step 558, the processing unit 110 may consider additional sources of information to further develop a safety model for the vehicle 200 in the context of its environment. The processing unit 110 may use the safety model to define a context in which the system 100 may safely execute autonomous control of the vehicle 200. To develop the safety model, in some embodiments, the processing unit 110 may take into account the position and movement of other vehicles, the detected road edges and obstacles, and / or general road shape descriptions extracted from map data (such as data from the map database 160). By considering additional information sources, the processing unit 110 may provide redundancy for detecting pavement marking and lane geometry and increase the reliability of the system 100.FIG. 5D is a flow diagram illustrating an example process 500D for detecting traffic lights in a set of images, in accordance with the disclosed embodiments. The processing unit 110 may execute the monocular image analysis module 402 to implement the process 500D. At step 560, the processing unit 110 may scan the set of images and identify objects that appear at locations in the images that are likely to contain traffic lights. For example, the processing unit 110 may filter the identified objects to create a set of candidate objects excluding those objects likely not to correspond to the traffic lights. The filtering may be based on various characteristics associated with traffic lights, such as shape, dimensions, texture, position (e.g., relative to the vehicle 200), and the like. Such characteristics may be based on several examples of traffic lights and traffic control signals and stored in a database. In some embodiments, the processing unit 110 may perform multiframe analysis on the set of candidate objects reflecting possible traffic lights. For example, the processing unit 110 may track the candidate objects over successive frames, estimate the real position of the candidate objects, and filter out those objects that are moving (which are likely not a traffic light). In some embodiments, the processing unit 110 may perform color analysis on the candidate objects and identify the relative position of the detected colors that appear within possible traffic lights.At step 562, the processing unit 110 may analyze the geometry of an intersection. The analysis may be made based on any combination of the following elements: (i) the number of lanes detected on both sides of the vehicle 200, (ii) markings (such as arrow markings) detected on the road, and (iii) descriptions of the intersection extracted from map data (such as data from the map database 160). The processing unit 110 may perform the analysis using information derived from the execution of the monocular analysis module 402. In addition, the processing unit 110 may determine a match between the traffic lights detected at step 560 and the lanes appearing in the vicinity of the vehicle 200.As the vehicle 200 approaches the intersection, the processing unit 110 may update the confidence level associated with the analyzed intersection geometry and detected traffic lights at step 564. For example, the number of traffic lights estimated to appear at the intersection compared to the number actually appearing at the intersection may affect the confidence level. Thus, based on the confidence level, the processing unit 110 may delegate control to the driver of the vehicle 200 to improve the safety conditions. By performing steps 560, 562, and 564, the processing unit 110 may identify traffic lights that appear within the set of captured images and analyze intersection geometry information. Based on the identification and analysis, the processing unit 110 may cause one or more navigation responses in the vehicle 200 as described above in connection with FIG. 5A.FIG. 5E is a flowchart illustrating an example process 500E for causing one or more navigation responses in the vehicle 200 based on a vehicle path, in accordance with the disclosed embodiments. At step 570, the processing unit 110 may construct an initial vehicle path associated with the vehicle 200. The vehicle path may be represented using a set of points expressed in coordinates (x, z), and the distance d i between two points in the set of points may fall within the range of 1 to 5 meters. In one embodiment, the processing unit 110 may construct the initial vehicle path using two polynomials, such as left and right road polynomials. The processing unit 110 may calculate the geometric center between the two polynomials and offset each point included in the resulting vehicle path by a predetermined offset (e.g., an offset of a smart lane), if any (an offset of zero may correspond to a trip in the center of a lane). The offset may be in a direction perpendicular to a segment between any two points in the vehicle path. In another embodiment, the processing unit 110 may use a polynomial and an estimated lane width to offset each point of the vehicle path by half the estimated lane width plus a predetermined offset (e.g., an offset of a smart lane).At step 572, the processing unit 110 may update the vehicle path constructed at step 570. The processing unit 110 may reconstruct the vehicle path constructed at step 570 using a higher resolution such that the distance d k between two points in the set of points representing the vehicle path is less than the distance d i described above. For example, the distance d k may fall within the range of 0.1 to 0.3 meters. The processing unit 110 may reconstruct the vehicle path using a parabolic spline algorithm that may provide a cumulative distance vector S corresponding to the total length of the vehicle path (i.e., based on the set of points representing the vehicle path).At step 574, the processing unit 110 may determine a look-ahead point (expressed in coordinates as (x l, z l)) based on the updated vehicle path constructed at step 572. The processing unit 110 may extract the look-ahead point from the cumulative distance vector S, and the look-ahead point may be associated with a look-ahead distance and look-ahead time. The look-ahead distance, which may have a lower limit in the range of 10 to 20 meters, may be calculated as the product of the speed of the vehicle 200 and the look-ahead time. For example, if the speed of the vehicle 200 decreases, the look-ahead distance may also decrease (e.g., until it reaches the lower limit). The look-ahead time, which may be in the range of 0.5 to 1.5 seconds, may be inversely proportional to the gain of one or more control loops associated with causing a navigation response in the vehicle 200, such as the control loop for heading error tracking. For example, the gain of the heading error tracking control loop may depend on the bandwidth of a yaw rate loop, a steering actuator loop, vehicle lateral dynamics, and the like. Thus, the higher the gain of the control loop for heading error tracking, the less the look-ahead time.At step 576, the processing unit 110 may determine a heading error and a yaw rate command based on the viewpoint determined at step 574. The processing unit 110 may determine the heading error by calculating the arctanof the look-ahead point, e.g., arctan(x l / z l). The processing unit 110 may determine the yaw rate command as a product of the heading error and a high control gain. The high control gain may be equal to: (2 / look-ahead time) when the look-ahead distance is not at the lower limit. Otherwise, the high control gain may be equal to: (2* speed of the vehicle 200 / look ahead distance).FIG. 5F is a flowchart illustrating an example process 500F for determining whether a preceding vehicle is changing lanes, in accordance with the disclosed embodiments. At step 580, the processing unit 110 may determine navigation information associated with a preceding vehicle (e.g., a vehicle traveling in front of the vehicle 200). For example, the processing unit 110 may determine the position, speed (e.g., direction and speed), and / or acceleration of the preceding vehicle using the techniques described above in connection with FIGS. 5A and 5B. The processing unit 110 may also determine one or more road polynomials, a look-ahead point (associated with the vehicle 200), and / or a spiral lane (e.g., a set of points describing a path taken by the preceding vehicle) using the techniques described above in connection with FIG. 5E.At step 582, the processing unit 110 may analyze the navigation information determined at step 580. In one embodiment, the processing unit 110 may calculate the distance between a scroll lane and a road polynomial (e.g., along the lane). When the variance of this distance along the lane exceeds a predetermined threshold (for example, 0.1 to 0.2 meters on a straight road, 0.3 to 0.4 meters on a moderately curved road, and 0.5 to 0.6 meters on a sharp corner road), the processing unit 110 may determine that the preceding vehicle is likely to change lanes. In the case that multiple vehicles are detected traveling in front of the vehicle 200, the processing unit 110 may compare the scroll lanes associated with each vehicle. Based on the comparison, the processing unit 110 may determine that a vehicle whose scroll lane does not match the scroll lanes of the other vehicles is likely to change lanes. The processing unit 110 may additionally compare the curvature of the auger lane (associated with the preceding vehicle) with the expected curvature of the road segment in which the preceding vehicle is traveling. The expected curvature may be extracted from map data (e.g., data from the map database 160), road polynomials, other vehicle auger lanes, pre-knowledge of the road, and the like. When the difference in curvature of the spiral lane and the expected curvature of the road segment exceeds a predetermined threshold, the processing unit 110 may determine that the preceding vehicle is likely to change lanes.In another embodiment, the processing unit 110 may compare the current position of the preceding vehicle with the viewpoint (associated with the vehicle 200) over a specific period of time (e.g., 0.5 to 1.5 seconds). When the distance between the current position of the preceding vehicle and the viewpoint varies during the specific period and the cumulative sum of the variation exceeds a predetermined threshold (for example, 0.3 to 0.4 meters on a straight road, 0.7 to 0.8 meters on a moderately curved road, and 1.3 to 1.7 meters on a sharp corner road), the processing unit 110 may determine that the preceding vehicle is likely to change lanes. In another embodiment, the processing unit 110 may analyze the geometry of the worm track by comparing the lateral distance travelled along the track to the expected curvature of the worm track. The expected radius of curvature can be determined by the following calculation: (δ z2+ δ x2) / 2 / ( δ x), where δ x represents the lateral distance travelled and δ z represents the longitudinal distance travelled. When the difference between the lateral distance traveled and the expected curvature exceeds a predetermined threshold (e.g., 500 to 700 meters), the processing unit 110 may determine that the preceding vehicle is likely to change lanes. In another embodiment, the processing unit 110 may analyze the position of the preceding vehicle. If the position of the preceding vehicle obscures a road polynomial (e.g., the preceding vehicle is superimposed on the road polynomial), then the processing unit 110 may determine that the preceding vehicle is likely to change lanes. In the case where the position of the preceding vehicle is such that another vehicle is detected in front of the preceding vehicle and the spiral lanes of the two vehicles are not parallel, the processing unit 110 may determine that the (closer) preceding vehicle is likely to change lanes.At step 584, the processing unit 110 may determine whether or not the preceding vehicle 200 is changing lanes based on the analysis performed at step 582. For example, the processing unit 110 may make the determination based on a weighted average of the individual analyses performed at step 582. Under such a scheme, for example, a value of "1" (and "0" may be assigned to a decision by the processing unit 110 that the preceding vehicle is likely to change lanes based on a certain type of analysis to represent a determination that the preceding vehicle is likely not to change lanes). Different analyses performed at step 582 may be assigned different weights, and the disclosed embodiments are not limited to any particular combination of analyses and weights.FIG. 6 is a flow diagram illustrating an example process 600 for causing one or more navigation responses based on stereo image analysis, in accordance with the disclosed embodiments. At step 610, the processing unit 110 may receive a first and second plurality of images via the data interface 128. For example, cameras included in the image capture unit 120 (such as the image capture devices 122 and 124 with the fields of view 202 and 204) may capture first and second pluralities of images of an area in front of the vehicle 200 and transmit them to the processing unit 110 via a digital connection (e.g., USB, wireless, Bluetooth, etc.). In some embodiments, the processing unit 110 may receive the first and second pluralities of images via two or more data interfaces. The disclosed embodiments are not limited to particular data interface configurations or protocols.At step 620, the processing unit 110 may execute the stereo image analysis module 404 to perform stereo image analysis of the first and second pluralities of images to generate a 3D map of the road in front of the vehicle and recognize features within the images, such as lane markings, vehicles, pedestrians, road signs, highway exit ramps, traffic lights, road hazards, and the like. Stereo image analysis may be performed in a manner similar to the steps described above in connection with Figures 5A-5D. For example, the processing unit 110 may execute the stereo image analysis module 404 to recognize candidate objects (e.g., vehicles, pedestrians, road markings, traffic lights, road hazards, etc.) within the first and second pluralities of images, filter out a subset of the candidate objects based on various criteria, and perform multiframe analysis, make measurements, and determine a confidence level for the remaining candidate objects. In performing the above steps, the processing unit 110 may consider information of both the first and second pluralities of images and information of a set of images alone. For example, the processing unit 110 may analyze the differences in pixel plane data (or other data subsets from the two streams of captured images) for a candidate object that appears in both the first and second pluralities of images. As another example, the processing unit 110 may estimate a position and / or speed of a candidate object (e.g., relative to the vehicle 200) by observing that the object appears in one of the plurality of images, but not in the other or relative to other differences that may exist relative to objects that appear when the two image streams. For example, position, velocity, and / or acceleration relative to the vehicle 200 may be determined based on trajectories, positions, motion features, etc., of features associated with an object appearing in one or both of the image streams.At step 630, the processing unit 110 may execute the navigation response module 408 to generate one or more navigation responses in the vehicle 200 based on the analysis performed at step 620 and the techniques described above in connection with FIG. 4. Navigation responses may include, for example, a turn, a lane shift, a change in acceleration, a change in speed, braking, and the like. In some embodiments, the processing unit 110 may use data derived from execution of the speed and acceleration module 406 to cause the one or more navigation responses. In addition, multiple navigation reactions may occur simultaneously, sequentially, or in any combination thereof.FIG. 7 is a flow diagram illustrating an example process 700 for causing one or more navigation responses based on analysis of three sets of images, in accordance with disclosed embodiments. At step 710, the processing unit 110 may receive a first, second, and third plurality of images via the data interface 128. For example, cameras included in the image capture unit 120 (such as the image capture devices 122, 124, and 126 having the fields of view 202, 204, and 206) may capture first, second, and third pluralities of images of an area in front of and / or to the side of the vehicle 200 and transmit them to the processing unit 110 via a digital connection (e.g., USB, wireless, Bluetooth, etc.). In some embodiments, the processing unit 110 may receive the first, second, and third pluralities of images via three or more data interfaces. For example, each of the image capture devices 122, 124, 126 may have an associated data interface for communicating data to the processing unit 110. The disclosed embodiments are not limited to particular data interface configurations or protocols.At step 720, the processing unit 110 may analyze the first, second, and third pluralities of images to identify features within the images, such as lane markings, vehicles, pedestrians, road signs, highway exit ramps, traffic lights, road hazards, and the like. The analysis can be performed in a manner similar to the steps described above in connection with Figures 5A-5D and 6. For example, the processing unit 110 may perform monocular image analysis (e.g., via execution of the monocular image analysis module 402 and based on the steps described above in connection with FIGS. 5A-5D ) on each of the first, second, and third pluralities of images. Alternatively, the processing unit 110 may perform stereo image analysis (e.g., via execution of the stereo image analysis module 404 and based on the steps described above in connection with FIG. 6 ) on the first and second pluralities of images, the second and third pluralities of images, and / or the first and third pluralities of images. The processed information corresponding to the analysis of the first, second and / or third plurality of images may be combined. In some embodiments, the processing unit 110 may perform a combination of monocular and stereo image analyses. For example, the processing unit 110 may perform monocular image analysis (e.g., via execution of the monocular image analysis module 402) on the first plurality of images and stereo image analysis (e.g., via execution of the stereo image analysis module 404) on the second and third plurality of images. The configuration of the image capture devices 122, 124, and 126, including their respective locations and fields of view 202, 204, and 206, may affect the types of analyses performed on the first, second, and third pluralities of images. The disclosed embodiments are not limited to any particular configuration of the image capture devices 122, 124, and 126 or the types of analysis performed on the first, second, and third pluralities of images.In some embodiments, processing unit 110 may perform testing on system 100 based on the images captured and analyzed at steps 710 and 720. Such testing may provide an indicator of the overall performance of the system 100 for particular configurations of the image capture devices 122, 124, and 126. For example, the processing unit 110 may determine the proportion of the "false hits" (e.g., cases where the system 100 has incorrectly determined the presence of a vehicle or pedestrian) and the "misses.".At step 730, the processing unit 110 may cause one or more navigation responses in the vehicle 200 based on information derived from two of the first, second, and third pluralities of images. The selection of two of the first, second, and third pluralities of images may depend on various factors, such as the number, types, and sizes of objects detected in each of the plurality of images. The processing unit 110 may also make the selection based on image quality and resolution, the effective field of view reflected in the images, the number of frames captured, the extent to which one or more objects of interest actually appear in the frames (e.g., the percentage of frames in which an object appears, the portion of the object appearing in each of these frames, etc.), and the like.In some embodiments, processing unit 110 may select information derived from two of the first, second, and third pluralities of images by determining the extent to which information derived from one image source is consistent with information derived from other image sources. For example, the processing unit 110 may combine the processed information derived from each of the image capture devices 122, 124, and 126 (whether by monocular analysis, stereo analysis, or any combination of the two) and determine visual indicators (e.g., lane markings, a detected vehicle and its location and / or path, a detected traffic light, etc.) consistent across the images captured by each of the image capture devices 122, 124, and 126. The processing unit 110 may also exclude information inconsistent with the captured images (e.g., a vehicle changing lanes, a lane model indicating a vehicle too close to the vehicle 200, etc.). Thus, the processing unit 110 may select information derived from two of the first, second, and third pluralities of images based on the determinations of consistent and inconsistent information.Navigation responses may include, for example, a turn, a lane shift, a change in acceleration, and the like. The processing unit 110 may cause the one or more navigation responses based on the analysis performed at step 720 and the techniques described above in connection with FIG. 4. The processing unit 110 may also use data derived from execution of the speed and acceleration module 406 to cause the one or more navigation responses. In some embodiments, the processing unit 110 may cause the one or more navigation responses based on a relative position, relative speed, and / or relative acceleration between the vehicle 200 and an object detected within any of the first, second, and third pluralities of images. Multiple navigation reactions may occur simultaneously, sequentially, or in any combination thereof.Sparse Road Model for Autonomous Vehicle NavigationIn some embodiments, the disclosed systems and methods may use a sparse map for autonomous vehicle navigation. In particular, the sparse map may be for autonomous vehicle navigation along a road segment. For example, the sparse map may provide sufficient information to navigate an autonomous vehicle without storing and / or updating a large amount of data. As discussed in more detail below, an autonomous vehicle may use the sparse map to navigate one or more roads based on one or more stored trajectories.Sparse Map for Autonomous Vehicle NavigationIn some embodiments, the disclosed systems and methods may generate a sparse map for autonomous vehicle navigation. For example, the sparse map may provide sufficient information for navigation without requiring excessive data storage or data transfer rates. As discussed in more detail below, a vehicle (which may be an autonomous vehicle) may use the sparse map to navigate one or more roads. For example, in some embodiments, the sparse map may include data relating to a road and potentially landmarks along the road that may be sufficient for vehicle navigation, but also have small data prints. For example, sparse data maps, described in more detail below, may require significantly less memory space and data transmission bandwidth compared to digital maps that include detailed map information, such as image data collected along a road.For example, rather than storing detailed representations of a road segment, the sparse data map may store three-dimensional polynomial representations of preferred vehicle paths along a road. These ways may require very little data storage space. Further, landmarks may be identified in the sparse data maps described and included in the street model of the sparse map to aid in navigation. These landmarks may be at any distance suitable for enabling vehicle navigation, but in some cases such landmarks need not be identified and incorporated into the model with high densities and short distances. Rather, navigation may be possible in some cases based on landmarks that are at least 50 meters, at least 100 meters, at least 500 meters, at least 1 kilometer, or at least 2 kilometer apart. As discussed in more detail in other sections, the sparse map may be generated based on data collected or measured from vehicles equipped with various sensors and devices, such as image capture devices, global positioning system sensors, motion sensors, etc., as the vehicles travel along lanes. In some cases, the sparse map may be generated based on data collected during multiple trips of one or more vehicles along a particular roadway. Generating a sparse map using multiple trips of one or more vehicles may be referred to as "crowdsourcing" a sparse map.In accordance with the disclosed embodiments, an autonomous vehicle system may use a sparse map for navigation. For example, the disclosed systems and methods may distribute a sparse map for generating a road navigation model for an autonomous vehicle and may navigate an autonomous vehicle along a road segment using a sparse map and / or a generated road navigation model. Sparse maps in accordance with the present disclosure may include one or more three-dimensional contours that may represent predetermined trajectories that autonomous vehicles may cross as they move along associated road segments.Sparse maps in accordance with the present disclosure may also include data representing one or more road features. Such road features may include recognized landmarks, road signature profiles, and any other road-related features useful in navigating a vehicle. Sparse maps in accordance with the present disclosure may enable autonomous navigation of a vehicle based on relatively small amounts of data included in the sparse map. For example, rather than including detailed representations of a road, such as road edges, road curvature, images associated with road segments, or data detailing other physical features associated with a road segment, the disclosed embodiments of the sparse map may require relatively little storage space (and relatively little bandwidth when transferring portions of the sparse map to a vehicle), but may still adequately provide autonomous vehicle navigation. The small data footprint of the disclosed sparse maps, discussed in more detail below, may be achieved in some embodiments by storing representations of road-related items that require small amounts of data, but still enable autonomous navigation.For example, rather than storing detailed representations of various aspects of a road, the disclosed sparse maps may store polynomial representations of one or more trajectories that a vehicle may follow along the road. Thus, rather than having to store (or transmit) details regarding the physical nature of the road to enable navigation along the road, a vehicle may be navigated along a particular road segment using the disclosed sparse maps without having to interpret physical aspects of the road in some cases, but rather by aligning its travel path with a trajectory (e.g., a polynomial spline) along the particular road segment. In this way, the vehicle may be navigated primarily based on the stored trajectory (e.g., a polynomial spline), which may require much less memory space than an approach that includes storing roadway images, road parameters, road layout, etc.In addition to the stored polynomial representations of trajectories along a road segment, the disclosed sparse maps may also include small data objects that may represent a road feature. In some embodiments, the small data objects may include digital signatures derived from a digital image (or signal) obtained from a sensor (e.g., a camera or other sensor, such as a suspension sensor) onboard a vehicle traveling along the road segment. The digital signature may have a reduced size relative to the signal detected by the sensor. In some embodiments, the digital signature may be generated to be compatible with a classifier function configured to recognize and identify the road feature from the signal detected by the sensor during a subsequent trip, for example. In some embodiments, a digital signature may be generated such that the digital signature has a footprint that is as small as possible while maintaining the ability to correlate or match the road feature with the stored signature based on an image (or digital signal generated by a sensor when the stored signature is not based on an image and / or includes other data) of the road feature captured by a camera onboard a vehicle traveling along the same road segment at a subsequent time.In some embodiments, a size of the data objects may be further associated with uniqueness of the road feature. For example, for a road feature detectable by a camera onboard a vehicle, and when the camera system onboard the vehicle is coupled to a classifier capable of distinguishing the image data corresponding to that road feature as being associated with a particular type of road feature, for example a road sign, and when such a road sign is locally unique in that area (e.g., there is no identical road sign or road sign of the same type in proximity), it may be sufficient to store data indicative of the type of road feature and its location.As will be discussed in more detail below, road features (e.g., landmarks along a road segment) may be stored as small data objects that may represent a road feature in relatively few bytes, while at the same time providing sufficient information for recognizing and using such a feature for navigation. In one example, a road sign may be identified as a recognized landmark on which navigation of a vehicle may be based. A representation of the road sign may be stored in the sparse map to include, for example, some bytes of data indicating a type of landmark (e.g., a stop sign) and some bytes of data indicating a location of the landmark (e.g., coordinates). Navigating based on such data light representations of the landmarks (e.g., using representations sufficient to locate, recognize, and navigate based on the landmarks) may provide a desired level of navigation functionality associated with sparse maps without significantly increasing the data overhead associated with the sparse maps. This sparse representation of landmarks (and other road features) may leverage the sensors and processors included onboard such vehicles configured to detect, identify, and / or classify certain road features.For example, if a character or even a particular type of character is locally unique in a given area (e.g., if there is no other character or character of the same type), the sparse map may use data indicating a type of landmark (a character or a particular type of character) and during navigation (e.g., autonomous navigation), if a camera onboard an autonomous vehicle captures an image of the area that includes a character (or a particular type of character), the processor may process the image, detect the character (if it is actually present in the image), classify the image as a character (or as a particular type of character), and classify the location of the image with the location of the character, as stored in the sparse map.The sparse map may include any suitable representation of the objects identified along a road segment. In some cases, the objects may be referred to as semantic objects or non-semantic objects. Semantic objects may be, for example, objects associated with a predetermined type classification. This type classification may be useful to reduce the amount of data required to describe the semantic object detected in an environment, which may be beneficial in both the gathering phase (e.g., to reduce the cost associated with bandwidth usage for transmitting driving information from multiple gathering vehicles to a server) and the navigation phase (e.g., reducing map data may accelerate the transmission of map tiles from a server to a navigating vehicle and also reduce the cost associated with bandwidth usage for such transmissions). Semantic object classification types may be assigned to any type of objects or features found along a roadway.Semantic objects may be further divided into two or more logical groups. For example, in some cases, a set of semantic object types may be associated with predetermined dimensions. Such semantic objects may be certain speed limit signs, heading signs, merging signs, stop signs, traffic lights, direction arrows on the roadway, manhole covers, or any other type of objects that may be associated with a standardized size. An advantage of such semantic objects is that only very little data is needed to represent / completely define the objects. For example, if the standardized size of a speed limit sign is known, a collection vehicle needs to identify (by analyzing a captured image) only the presence of a speed limit sign (of a recognized type) along with an indication of a position of the recognized speed limit sign (e.g., a 2D position in the captured image (or, alternatively, a 3D position in real coordinates) of a center of the sign or a particular corner of the sign) to provide sufficient information for map generation on the server side. When 2D image positions are transmitted to the server, a position associated with the captured image at which the sign was detected may also be transmitted, in order for the server to determine a real position of the sign (e.g., by structure-in-motion techniques using multiple captured images from one or more collection vehicles). Even with this limited information (requiring only a few bytes to define each detected object), the server can create the map including a fully presented speed limit sign based on the type classification (representative of a speed limit sign) received from one or more collection vehicles along with the position information for the detected sign.Semantic objects may also include other recognized object or feature types that are not associated with certain standardized features. Such objects or features may be tapped holes, tar seams, light poles, non-standardized signs, curbs, trees, branches, or other detected object types having one or more variable features (e.g., variable dimensions). In such cases, in addition to transmitting an indication of the detected object or feature type (e.g., tap hole, post, etc.) and the position information for the detected object or feature to a server, a collection vehicle may also transmit an indication of the size of the object or feature. The size may be expressed in 2D image dimensions (e.g., with a bounding box or one or more dimension values) or in real dimensions (determined by structure-in-motion computations based on the results of LIDAR or RADAR systems, based on the outputs of trained neural networks, etc.).Non-semantic objects or features may include any detectable objects or features that do not fall within an accepted category or type, but yet may still provide valuable information for map generation. In some cases, such non-semantic features may be a detected corner of a building or a corner of a detected window of a building, a unique stone or object near a roadway, a concrete chip on a roadway edge, or any other detectable object or feature. Upon detecting such an object or feature, one or more collection vehicles may transmit the position of one or more points (2D pixels or 3D points in the real world) associated with the detected object / feature to a map creation server. Additionally, a compressed or simplified image segment (e.g., an image hash) may be generated for a portion of the captured image that includes the detected object or feature. This image hash may be calculated based on a predetermined image processing algorithm and may form an effective signature for the detected non-semantic object or feature. Such a signature may be useful for navigation with respect to a sparse map that includes the non-semantic feature or object, as a vehicle traversing the roadway may apply an algorithm similar to the algorithm used to generate the image hash to confirm / verify the presence of the mapped non-semantic feature or object in a captured image. Using this technique, non-semantic features may contribute to richness of sparse maps (e.g., to improve their usefulness in navigation) without adding significant data overhead.As noted, target trajectories may be stored in the sparse map. These target trajectories (e.g., 3D splines) may represent the preferred or recommended paths for each available lane of a roadway, each valid footway through an intersection, for merges and exits, etc. In addition to the target trajectories, other road features may also be detected, collected and included in the sparse maps in the form of representative splines. Such features may be, for example, road edges, lane markings, curbs, guardrails, or other objects or features extending along a roadway or road segment.Generating a sparse cardIn some embodiments, a sparse map may include at least one line representation of a feature of the road surface extending along a road segment and a plurality of landmarks associated with the road segment. In certain aspects, the sparse map may be generated by crowdsourcing, for example, by image analysis of a plurality of images captured as one or more vehicles cross the road segment.FIG. 8 shows sparse map 800 that may be accessed by one or more vehicles, e.g., vehicle 200 (which may be an autonomous vehicle), to provide autonomous vehicle navigation. Sparse map 800 may be stored in a memory, such as memory 140 or 150. Such storage devices may include any type of non-transitory storage devices or computer readable media. For example, in some embodiments, the memory 140 or 150 may include hard drives, CDs, flash memory, magnetic storage devices, optical storage devices, etc. In some embodiments, the low density map 800 may be stored in a database (e.g., map database 160), which may be stored in a memory 140 or 150 or other types of storage devices.In some embodiments, sparse map 800 may be stored on a storage device or a non-transitory computer readable medium provided onboard vehicle 200 (e.g., a storage device included in a navigation system onboard vehicle 200). A processor (e.g., processing unit 110) provided on the vehicle 200 may access the sparse map 800 stored in the storage device or the computer readable medium provided onboard the vehicle 200 to generate navigation instructions for guiding the autonomous vehicle 200 when the vehicle crosses a road segment.However, sparse map 800 does not need to be stored locally with respect to a vehicle. In some embodiments, sparse map 800 may be stored on a storage device or computer readable medium provided on a remote server that communicates with vehicle 200 or a device associated with vehicle 200. A processor (e.g., processing unit 110) provided on the vehicle 200 may receive data included in the sparse map 800 from the remote server and may execute the data for guiding autonomous driving of the vehicle 200. In such embodiments, the remote server may store all or a portion of sparse map 800. Accordingly, the storage device or computer readable medium provided onboard the vehicle 200 and / or onboard one or more additional vehicles may store the remaining portion(s) of the sparse map 800.Moreover, in such embodiments, sparse map 800 may be made accessible to a plurality of vehicles crossing different road segments (e.g., dozens, hundreds, thousands or millions of vehicles, etc.). It should also be noted that sparse map 800 may include multiple sub-maps. For example, in some embodiments, sparse map 800 may include hundreds, thousands, millions, or more sub-maps (e.g., map tiles) that may be used in navigating a vehicle. Such sub-maps may be referred to as local maps or map tiles, and a vehicle traveling along a roadway may access any number of local maps relevant to the location at which the vehicle is traveling. The local map sections of sparse map 800 may be stored with a global navigation satellite system (GNSS) key as an index for sparse map database 800. While the calculation of steering angles for navigating a host vehicle may be performed in the present system without dependence on a GNSS position of the host vehicle, road features or landmarks, such GNSS information may be used to retrieve relevant local maps.In general, sparse map 800 may be generated based on data collected from one or more vehicles as they travel along the lanes. For example, using sensors onboard the one or more vehicles (e.g., cameras, speedometers, GPS, accelerometers, etc.), the trajectories that the one or more vehicles travel along a roadway may be recorded, and the polynomial representations of a preferred trajectory for vehicles that take subsequent trips along the roadway may be determined based on the collected trajectories that are driven by the one or more vehicles. Similarly, data collected by the one or more vehicles may aid in identifying potential landmarks along a particular roadway. The data collected by the passing vehicles may also be used to identify road profile information, such as road width profiles, road unevenness profiles, line distance profiles, condition of the road, etc. The collected information may be used to create and distribute a low density map 800 (e.g., for local storage or by instant data transfer) that may be used to navigate one or more autonomous vehicles. However, in some embodiments, map generation may not end at the initial generation of the map. As will be discussed in more detail below, sparse map 800 may be continuously or periodically updated based on data collected from vehicles as those vehicles continue to cross lanes included in sparse map 800.Data recorded in sparse map 800 may include position information based on global positioning system (GPS) data. For example, location information in the sparse map 800 may be included for various map elements, including, for example, landmarks, road profile locations, etc. Locations for map elements included in the sparse map 800 may be obtained using GPS data collected from vehicles traversing a road. For example, a vehicle passing an identified landmark may determine a location of the identified landmark using GPS position information associated with the vehicle and a determination of a location of the identified landmark relative to the vehicle (e.g., based on image analysis of data collected from one or more cameras onboard the vehicle). Such location determinations of an identified landmark (or any other feature included in sparse map 800) may be repeated as additional vehicles pass the location of the identified landmark. Some or all of the additional location determinations may be used to refine the location information stored in sparse map 800 relative to the identified landmark. For example, in some embodiments, multiple position measurements relative to a particular feature stored in sparse map 800 may be averaged together. However, any other mathematical operations may also be used to refine a stored location of a map element based on a plurality of determined locations for the map element.In a particular example, the collection vehicles may cross a particular road segment. Each collection vehicle captures images of its respective environments. The images may be collected at any suitable image capture rate (e.g., 9 Hz, etc.). (An) image analysis processor(s) on board each collection vehicle analyzes (analyze) the captured images to detect the presence of semantic and / or non-semantic features / objects. At a high level, the collection vehicles transmit information regarding the detections of the semantic and / or non-semantic objects / features to a mapping server along with the locations associated with those objects / features. In more detail, type indicators, dimension indicators, etc. may be transmitted along with the position information. The location information may include any suitable information that enables the mapping server to aggregate the detected objects / features into a sparse map useful in navigation. In some cases, the position information may include one or more 2D image positions (e.g., X-Y pixel positions) in a captured image in which the semantic or non-semantic features / objects have been detected. Such image positions may correspond to a center of the feature / object, a corner, etc. In this scenario, to assist the mapping server in reconstructing the driving information and aligning the driving information from multiple collection vehicles, each collection vehicle may also provide the server with a location (e.g., a GPS location) at which each image was captured.In other cases, the collection vehicle may provide the server with one or more real 3D points associated with the detected objects / features. Such 3D points may refer to a predetermined origin (such as the origin of a driving segment) and may be determined using any suitable technique. In some cases, a structure-in-motion technique may be used to determine the real 3D position of a detected object / feature. For example, a particular object, such as a particular speed limit sign, may be detected in two or more captured images. Using information such as the known own motion (speed, trajectory, GPS position, etc.) of the collection vehicle between the captured images, along with observed changes in the speed limit sign in the captured images (change in X-Y pixel position, change in size, etc.), the real position of one or more points associated with the speed limit sign can be determined and forwarded to the mapping server. Such an approach is optional because it requires a higher computational effort for the systems of the collection vehicle. The sparse map of the disclosed embodiments may enable autonomous navigation of a vehicle using relatively small amounts of stored data. In some embodiments, sparse map 800 may have a data density (e.g., including data representing target trajectories, landmarks, and any other stored road features) of less than 2 MB per kilometre roads, less than 1 MB per kilometre roads, less than 500 kB per kilometre roads, or less than 100 kB per kilometre roads. In some embodiments, the data density of sparse map 800 may be less than 10 kb per kilometer of roads, or even less than 2 kb per kilometer of roads (e.g., 1.6 kb per kilometer), or no more than 10 kb per kilometer of roads, or no more than 20 kb per kilometer of roads. In some embodiments, most, if not all, of the United States lanes may be autonomously navigated using a sparse map having a total of 4 GB or less data. These data density values may represent an average over an entire sparse map 800, over a local map within sparse map 800, and / or over a particular road segment within sparse map 800.As noted, sparse map 800 may include representations of a plurality of target trajectories 810 for guiding autonomous driving or navigation along a road segment. Such target trajectories can be stored as three-dimensional splines. The target trajectories stored in sparse map 800 may be determined based on, for example, two or more reconstructed trajectories of previous intersections of vehicles along a particular road segment. A road segment may be associated with a single target trajectory or multiple target trajectories. For example, on a two-lane road, a first target trajectory may be stored to represent an intended travel path along the road in a first direction, and a second target trajectory may be stored to represent an intended travel path along the road in another direction (e.g., opposite the first direction). Additional target trajectories may be stored with respect to a particular road segment. For example, a multi-lane road may store one or more target trajectories representing intended travel paths for vehicles in one or more lanes associated with the multi-lane road. In some embodiments, each lane of a multi-lane road may be associated with its own target trajectory. In other embodiments, fewer target trajectories may be stored than lanes present on a multi-lane road. In such cases, a vehicle that navigates the multi-lane road may use any of the stored target trajectories to guide its navigation by considering an amount of lane offset from a lane for which a target trajectory is stored (e.g., when a vehicle is traveling in the leftmost lane of a three-lane highway and a target trajectory is stored only for the mid lane of the highway, the vehicle may navigate using the mid-lane target trajectory by considering the amount of lane offset between the mid lane and the leftmost lane when generating navigation instructions).In some embodiments, the target trajectory may represent an ideal path that a vehicle should take when the vehicle is driving. The target trajectory may be located at an approximate center of a lane, for example. In other cases, the target trajectory may be elsewhere with respect to a road segment. For example, a trajectory may be approximately coincident with the center of a road, the edge of a road, or the edge of a lane, etc. In such cases, based on the trajectory of the target, the navigation may include a certain offset that needs to be maintained relative to the location of the trajectory of the target. Moreover, in some embodiments, the determined amount of offset to be maintained with respect to the location of the target trajectory may differ based on a vehicle type (e.g., a two-axis passenger vehicle may have a different offset than a more than two-axis truck along at least a portion of the target trajectory).Sparse map 800 may also include data relating to a plurality of predetermined landmarks 820 associated with particular road segments, local maps, etc. As will be discussed in more detail below, these landmarks may be used in the navigation of the autonomous vehicle. For example, in some embodiments, the landmarks may be used to determine a current position of the vehicle with respect to a stored target trajectory. With this position information, the autonomous vehicle may be capable of adjusting a direction of travel to correspond to a direction of the target trajectory at the particular location.The plurality of landmarks 820 may be identified and stored in the sparse map 800 at any suitable distance. In some embodiments, landmarks may be stored with relatively high densities (e.g., every few meters or more). However, in some embodiments, significantly larger landmark distance values may be used. For example, landmarks identified (or detected) in sparse map 800 may be 10 meters, 20 meters, 50 meters, 100 meters, 1 kilometer, or 2 kilometer apart. In some cases, the identified landmarks may be located at distances of even more than 2 kilometers from each other.Between landmarks and therefore between determinations of vehicle position relative to a target trajectory, the vehicle may navigate based on dead reckoning, where the vehicle uses sensors to determine its own motion and estimate its position relative to the target trajectory. As errors may accumulate during navigation through dead reckoning, position determinations relative to the target trajectory may become more and more inaccurate over time. The vehicle may use landmarks that appear in sparse map 800 (and their known locations) to eliminate dead reckoning induced errors in position determination. In this way, the identified landmarks included in sparse map 800 may serve as navigation anchors from which an accurate position of the vehicle relative to a target trajectory may be determined. Since some amount of error in location of the position may be acceptable, an identified landmark may not always be available to an autonomous vehicle. Rather, suitable navigation may also be possible based on landmark distances, as noted above, of 10 meters, 20 meters, 50 meters, 100 meters, 500 meters, 1 kilometer, 2 kilometer or more. In some embodiments, a density of 1 identified landmark every 1 km road may be sufficient to maintain longitudinal positioning accuracy within 1 m. Thus, not every potential landmark appearing along a road segment needs to be stored in sparse map 800.Moreover, in some embodiments, lane markings may be used to locate the vehicle during landmark distances. By using lane markings during landmark distances, accumulation of errors by dead reckoning during navigation may be minimized.In addition to target trajectories and identified landmarks, sparse map 800 may include information related to various other road features. For example, FIG. 9A illustrates a representation of curves along a particular road segment that may be stored in a low density map 800. In some embodiments, a single lane of a road may be modeled by a three-dimensional polynomial description of the left and right sides of the road. Such polynomials representing the left and right sides of a single lane are shown in FIG. 9A. Regardless of how many lanes a road may have, the road may be represented using polynomials in a similar manner as illustrated in FIG. 9A. For example, the left and right sides of a multi-lane road may be represented by polynomials similar to those in FIG. 9A, and the intermediate markings on a multi-lane road (e.g., dashed markings representing the boundaries of the lanes, solid yellow lines representing the boundaries between lanes traveling in different directions, etc.) may also be represented by polynomials as shown in FIG. 9A.As shown in FIG. 9A, a lane 900 may be represented using polynomials (e.g., first, second, third, or any suitable order polynomials). For illustrative purposes, lane 900 is shown as a two-dimensional lane and polynomials are shown as two-dimensional polynomials. As shown in FIG. 9A, the lane 900 includes a left side 910 and a right side 920. In some embodiments, more than one polynomial may be used to represent a location of each side of the road or lane boundary. For example, each of the left side 910 and the right side 920 may be represented by a plurality of polynomials of any suitable length. In some cases, the polynomials may have a length of about 100 m, although other lengths greater or less than 100 m may also be used. Additionally, the polynomials may overlap to facilitate seamless transitions in navigating based on subsequently encountered polynomials as a host vehicle travels along a roadway. For example, each of the left side 910 and the right side 920 may be represented by a plurality of third order polynomials separated into segments of about 100 meters in length (an example of the first predetermined range) and overlapping each other by about 50 meters. The polynomials representing the left side 910 and the right side 920 may or may not have the same order. For example, in some embodiments, some polynomials may be second-order polynomials, some third-order polynomials, and some fourth-order polynomials.In the example shown in FIG. 9A, the left side 910 of the lane 900 is represented by two groups of third order polynomials. The first group includes polynomial segments 911, 912, and 913. The second group includes polynomial segments 914, 915, and 916. The two groups, while substantially parallel to each other, follow the locations of their respective sides of the road. Polynomial segments 911, 912, 913, 914, 915, and 916 have a length of about 100 meters and overlap adjacent segments in the series by about 50 meters. However, as noted above, polynomials of different lengths and different overlap amounts may also be used. For example, the polynomials may have lengths of 500 m, 1 km, or more, and the overlap amount may vary from 0 to 50 m, 50 m to 100 m, or more than 100 m. Additionally, while FIG. 9A is shown as representing polynomials extending in 2D space (e.g., on the surface of the paper), it should be understood that these polynomials may represent curves extending in three dimensions (e.g., including a height component) to represent height changes in a road segment in addition to the X-Y curvature. In the example shown in FIG. 9A, the right side 920 of the lane 900 is further represented by a first group including polynomial segments 921, 922, and 923 and a second group including polynomial segments 924, 925, and 926.Returning to the trajectories of map 800, FIG. 9B shows a three-dimensional polynomial representing a trajectory for a vehicle moving along a particular road segment. The target trajectory represents not only the X-Y path that a host vehicle should travel along a particular road segment, but also the change in elevation that the host vehicle will experience when traveling along the road segment. Thus, each trajectory in the sparse map 800 may be represented by one or more three-dimensional polynomials, such as the three-dimensional polynomial 950 shown in FIG. 9B. Sparse map 800 may include a plurality of trajectories (e.g., millions or billions or more to represent trajectories of vehicles along different road segments along entire world lanes). In some embodiments, each target trajectory may correspond to a spline connecting three-dimensional polynomial segments.With respect to the data footprint of polynomial curves stored in sparse map 800, in some embodiments, each third degree polynomial may be represented by four parameters that each require four bytes of data. Suitable representations may be obtained with third degree polynomials requiring about 192 bytes of data per 100 meters. This may result in a data consumption / transfer of about 200 kb per hour for a host vehicle traveling at a speed of about 100 km / h.Sparse map 800 may describe the lane network using a combination of geometry descriptors and metadata. The geometry may be described by polynomials or splines as described above. The metadata may describe the number of lanes, specific characteristics (such as an autopool lane), and possibly other sparse designations. The total footprint of such indicators may be negligible.Accordingly, a sparse map according to embodiments of the present disclosure may include at least one line representation of a road surface feature extending along the road segment, each line representation representing a path along the road segment that substantially corresponds to the road surface feature. In some embodiments, as discussed above, the at least one line representation of the road surface feature may include a spline, polynomial representation, or curve. Moreover, in some embodiments, the road surface feature may include at least one of a road edge or a lane marking. Moreover, as discussed below with respect to "crowdsourced", the road surface feature may be identified by image analysis of a plurality of images captured as one or more vehicles cross the road segment.As noted above, sparse map 800 may include a plurality of predetermined landmarks associated with a road segment. For example, rather than storing actual images of landmarks and relying on image recognition analysis based on captured images and stored images, each landmark in sparse map 800 may be presented and recognized using less data than a stored actual image would require. Data representing landmarks may further include sufficient information to describe or identify the landmarks along a road. Storing data describing characteristics of landmarks instead of the actual images of landmarks may reduce the size of the sparse map 800.FIG. 10 illustrates examples of types of landmarks that may be displayed in the map 800. The landmarks may include any visible and identifiable objects along a road segment. The landmarks may be selected to be fixed and not change often with respect to their locations and / or content. The landmarks included in sparse map 800 may be useful in determining a location of vehicle 200 with respect to a target trajectory when the vehicle crosses a particular road segment. Examples of landmarks may include road signs, directional signs, general signs (e.g., rectangular signs), roadside attachments (e.g., lampposts, reflectors, etc.), and any other suitable category. In some embodiments, lane markings on the road may also be included as landmarks in sparse map 800.Examples of landmarks shown in FIG. 10 include traffic signs, directional signs, roadside attachments, and general signs. Traffic signs may include, for example, speed limit signs (e.g., speed limit sign 1000), forward travel signs (e.g., forward travel sign 1005), route number signs (e.g., route number sign 1010), traffic light signs (e.g., traffic light sign 1015), and stop signs (e.g., stop sign 1020). Directional signs may include a sign including one or more arrows indicating one or more directions to different locations. For example, directional signs may include a highway sign 1025 having arrows for steering vehicles to different roads or locations, an exit sign 1030 having an arrow for steering vehicles from a road, etc. Accordingly, at least one of the plurality of landmarks may include a road sign.General characters need not be traffic related. For example, general signs may include advertising panels used for advertising or a williams panel adjacent a boundary between two countries, states, districts, cities, or cities. Figure 10 shows a generic character 1040 ("Joe's Restaurant"). Although the general sign 1040 may have a rectangular shape as shown in FIG. 10, the general sign 1040 may also have other shapes such as square, circle, triangle, etc.Landmarks may also include roadside attachments. Roadside attachments may be objects that are not characters and need not be associated with traffic or directions. For example, roadside attachments may include lampposts (e.g., lamppost 1035), power lineposts, traffic lights, etc.Landmarks may also include beacons that may be specifically configured for use in an autonomous vehicle navigation system. For example, such beacons may include stand-alone structures placed at predetermined intervals to aid in navigating a host vehicle. Such beacons may also include visual / graphical information added to existing road signs (e.g., icons, emblems, bar codes, etc.) that may be identified or recognized by a vehicle traveling along a road segment. Such beacons may also include electronic components. In such embodiments, electronic beacons (e.g., RFID tags, etc.) may be used to transmit non-visual information to a host vehicle. Such information may include, for example, landmark identification and / or landmark location information that a host vehicle may use in determining its position along a target trajectory.In some embodiments, the landmarks included in sparse map 800 may be represented by a data object of a predetermined size. The data representing an landmark may include any suitable parameter for identifying a particular landmark. For example, in some embodiments, landmarks stored in sparse map 800 may include parameters such as a physical size of the landmark (e.g., to aid in estimating the distance to the landmark based on a known size / scale), a distance to a previous landmark, lateral offset, elevation, a type code (e.g., a landmark type - which type of direction signs, traffic signs, etc.), a GPS coordinate (e.g., to aid in global localization), and any other suitable parameters. Each parameter may be associated with a data size. For example, an landmark size may be stored using 8 bytes of data. A distance to a previous landmark, a lateral offset, and a height may be specified using 12 bytes of data. A type code associated with an landmark, such as a direction sign or a road sign, may require about 2 bytes of data. For general characters, an image signature that allows identification of the general character may be stored using 50 bytes of data storage. The landmark GPS location may be associated with 16 bytes of data storage. These data sizes for each parameter are only examples, and other data sizes may also be used. Presenting landmarks in sparse map 800 in this manner may provide a slender solution for efficiently presenting landmarks in the database. In some embodiments, objects may be referred to as default semantic objects and non-default semantic objects. A standard semantic object may include any class of objects for which there is a standardized set of properties (e.g., speed limit signs, warning signs, direction signs, traffic lights, etc., of known dimensions or other properties). A non-standard semantic object may include any object that is not associated with a standardized set of properties (e.g., general advertising signs, signs identifying business entities, potholes, trees, etc., which may have variable dimensions). Each non-standard semantic object may be represented with 38 bytes of data (e.g., 8 bytes for size; 12 bytes for distance to a previous landmark, lateral offset, and height; 2 bytes for a typecode; and 16 bytes for position coordinates). Standard semantic objects can be represented using even less data, as size information may not be needed by the mapping server to fully represent the object in the sparse map.Sparse map 800 may use a marking system to represent landmark types. In some cases, each traffic sign or direction sign may be associated with its own tag, which may be stored in the database as part of the landmark identification. For example, the database may include on the order of 1000 different markings to represent different traffic signs and on the order of about 10000 different markings to represent directional signs. Of course, any suitable number of markers may be used, and additional markers may be generated as desired. Multi-purpose characters may be represented with less than about 100 bytes (e.g., about 86 bytes, including 8 bytes for size; 12 bytes for distance to a previous landmark, lateral offset, and elevation; 50 bytes for an image signature; and 16 bytes for GPS coordinates), in some embodiments.Thus, for semantic road signs that do not require an image signature, the data density impact on sparse map 800 may be on the order of about 760 bytes per kilometer (e.g., 20 landmarks per km x 38 bytes per landmark=760 bytes), even at relatively high landmark densities of about 1 per 50 m. Even for general characters that include an image signature component, the data density impact is about 1.72 kB per km (e.g., 20 landmarks per km x 86 bytes per landmark=1.720 bytes). For semantic road signs, this corresponds to about 76 kb per hour of data usage for a vehicle traveling 100 km / h. For general signs, this corresponds to about 170 kb per hour for a vehicle traveling 100 km / h. It should be noted that in some environments (e.g., urban environments), there may be a much higher density of detected objects available for inclusion in the sparse map (perhaps more than one per meter). In some embodiments, a generally rectangular object, such as a rectangular character, in sparse map 800 may be represented by no more than 100 bytes of data. The representation of the generally rectangular object (e.g., the general character 1040) in the sparse map 800 may include a condensed image signature (e.g., the condensed image signature 1045) associated with the generally rectangular object. This condensed image signature can be used, for example, to aid in identifying a general character, for example, as a recognized landmark. Such condensed image signature (e.g., image information derived from actual image data representing an object) may avoid a need for storing an actual image of an object or a need for comparative image analysis performed on actual images to recognize landmarks.Referring to FIG. 10, the low density map 800 may include or store a compressed image signature 1045 associated with a generic character 1040, rather than an actual image of the generic character 1040. For example, after an image capture device (e.g., image capture device 122, 124, or 126) captures an image of the general character 1040, a processor (e.g., image processor 190 or any other processor that can process images either onboard or remotely relative to a host vehicle) can perform image analysis to extract / generate the condensed image signature 1045 that includes a unique signature or pattern associated with the general character 1040. In an embodiment, the condensed image signature 1045 may include a shape, color pattern, brightness pattern, or any other feature that may be extracted from the image of the general character 1040 to describe the general character 1040.For example, in FIG. 10, the circles, triangles, and stars shown in the compressed image signature 1045 may represent regions of different colors. The pattern represented by the circles, triangles, and stars may be stored in sparse map 800, e.g., within the 50 bytes designated to include an image signature. In particular, the circles, triangles, and stars are not necessarily intended to indicate that such shapes are stored as part of the image signature. Rather, these shapes are intended to conceptually represent recognizable regions with recognizable color differences, text regions, graphical shapes, or other variations of properties that may be associated with a general character. Such condensed image signatures can be used to identify an landmark in the form of a general character. For example, the condensed image signature may be used to perform a same-non-same analysis based on a comparison of a stored condensed image signature with image data captured using, for example, a camera onboard an autonomous vehicle.Accordingly, the plurality of landmarks may be identified by image analysis of the plurality of images captured when one or more vehicles cross the road segment. As discussed below with respect to "crowdsourcing", in some embodiments, the image analysis to identify the plurality of landmarks may include accepting potential landmarks when a ratio of images in which the landmark appears to images in which the landmark does not appear exceeds a threshold. Moreover, in some embodiments, the image analysis to identify the plurality of landmarks may include rejecting potential landmarks when a ratio of images in which the landmark does not appear to images in which the landmark appears exceeds a threshold.Returning to the trajectories that a host vehicle may use to navigate a particular road segment, FIG. 11A shows polynomial representations of trajectories captured during a process for constructing or maintaining a sparse map 800. A polynomial representation of a target trajectory included in sparse map 800 may be determined based on two or more reconstructed trajectories of previous intersections of vehicles along the same road segment. In some embodiments, the polynomial representation of the target trajectory included in sparse map 800 may be an aggregation of two or more reconstructed trajectories of previous traversals of vehicles along the same road segment. In some embodiments, the polynomial representation of the target trajectory included in sparse map 800 may be an average of the two or more reconstructed trajectories of previous traversals of vehicles along the same road segment. Other mathematical operations may also be used to construct a target trajectory along a roadway based on reconstructed trajectories collected from traversing vehicles along a roadway segment.As shown in FIG. 11A, a road segment 1100 may be traveled by a number of vehicles 200 at different times. Each vehicle 200 may collect data related to a path the vehicle has taken along the road segment. The distance traveled by a particular vehicle may be determined based on camera data, accelerometer information, speed sensor information, and / or GPS information, among other potential sources. Such data may be used to reconstruct trajectories of vehicles moving along the road segment, and based on these reconstructed trajectories, a target trajectory (or multiple target trajectories) may be determined for the particular road segment. Such target trajectories may represent a preferred path of a host vehicle (e.g., guided by an autonomous navigation system) as the vehicle travels along the road segment.In the example shown in FIG. 11A, a first reconstructed trajectory 1101 may be determined based on data received from a first vehicle traversing the road segment 1100 at a first time period (e.g., day 1), a second reconstructed trajectory 1102 may be obtained from a second vehicle traversing the road segment 1100 at a second time (e.g., day 2), and a third reconstructed trajectory 1103 may be obtained from a third vehicle traversing the road segment 1100 at a third time (e.g., day 3). Each trajectory 1101, 1102, and 1103 may be represented by a polynomial such as a three-dimensional polynomial. It should be noted that in some embodiments, each of the reconstructed trajectories may be composed aboard the vehicles traversing the road segment 1100.Additionally or alternatively, such reconstructed trajectories may be determined on a server side based on information received from vehicles traversing the road segment 1100. For example, in some embodiments, the vehicles 200 may transmit data to one or more servers related to their movement along the road segment 1100 (e.g., steering angle, heading, time, position, speed, sensed road geometry, and / or sensed landmarks, among others). The server may reconstruct trajectories for the vehicles 200 based on the received data. The server may also generate a target trajectory for guiding navigation of the autonomous vehicle traveling along the same road segment 1100 at a later time based on the first, second, and third trajectories 1101, 1102, and 1103. While a target trajectory may be associated with a single prior traversal of a road segment, in some embodiments, each target trajectory included in sparse map 800 may be determined based on two or more reconstructed trajectories of vehicles traversing the same road segment. In FIG. 11A, the trajectory of the target is represented by 1110. In some embodiments, the target trajectory 1110 may be generated based on an average of the first, second, and third trajectories 1101, 1102, and 1103. In some embodiments, the target trajectory 1110 included in the sparse map 800 may be an aggregation (e.g., a weighted combination) of two or more reconstructed trajectories.On the mapping server, the server may receive actual trajectories for a particular road segment from multiple collection vehicles traversing the road segment. To generate a target trajectory for each valid path along the road segment (e.g., each lane, each heading, each foot path through an intersection, etc.), the received actual trajectories may be aligned. The alignment process may include using detected objects / features identified along the road segment along with collected positions of those detected objects / features to correlate the actual collected trajectories with each other. Once aligned, an average or "best-fit" target trajectory for each available lane, etc., may be determined based on the aggregated, correlated / aligned actual trajectories.FIGS. 11B and 11C further illustrate the concept of target trajectories associated with road segments present within a geographic area 1111. As shown in FIG. 11B, a first road segment 1120 within a geographic region 1111 may include a multi-lane road that includes two lanes 1122 provided for vehicle travel in a first direction and two additional lanes 1124 provided for vehicle travel in a second direction opposite the first direction. The lanes 1122 and the lanes 1124 may be separated by a double yellow line 1123. The geographic area 1111 may also include a branch road segment 1130 that intersects with the road segment 1120. The road segment 1130 may include a two-lane road, each lane being dedicated to a different direction of travel. The geographic area 1111 may also include other road features, such as a stop line 1132, a stop sign 1134, a speed limit sign 1136, and a hazard sign 1138.As shown in FIG. 11C, sparse map 800 may include a local map 1140 that includes a road model to aid autonomous navigation of vehicles within geographic region 1111. For example, the local map 1140 may include target trajectories for one or more lanes associated with the road segments 1120 and / or 1130 within the geographic area 1111. For example, the local map 1140 may include target trajectories 1141 and / or 1142 that may be accessed or supported by an autonomous vehicle as it crosses lanes 1122. Similarly, local map 1140 may include target trajectories 1143 and / or 1144 that may be accessed or supported by an autonomous vehicle as it crosses lanes 1124. Further, the local map 1140 may include target trajectories 1145 and / or 1146 that may be accessed or supported by an autonomous vehicle as it crosses the road segment 1130. The target trajectory 1147 represents a preferred path that an autonomous vehicle should follow when transitioning from the lanes 1120 (and specifically relative to the target trajectory 1141 associated with a rightmost lane of the lanes 1120) to the road segment 1130 (and specifically relative to a target trajectory 1145 associated with a first side of the road segment 1130). Similarly, the target trajectory 1148 represents a preferred path that an autonomous vehicle should follow when transitioning from road segment 1130 (and specifically relative to the target trajectory 1146) to a portion of the road segment 1124 (and specifically, as shown, relative to a target trajectory 1143 associated with a left lane of the lanes 1124).Sparse map 800 may also include representations of other road-related features associated with geographic area 1111. For example, sparse map 800 may also include representations of one or more landmarks identified in geographic area 1111. Such landmarks may include a first landmark 1150 associated with the stop line 1132, a second landmark 1152 associated with the stop sign 1134, a third landmark associated with the speed limit sign 1154, and a fourth landmark 1156 associated with the hazard sign 1138. Such landmarks may be used, for example, to assist an autonomous vehicle in determining its current location with respect to one of the shown target trajectories, such that the vehicle may adjust its direction of travel to correspond to a direction of the target trajectory at the determined location.In some embodiments, sparse map 800 may also include road signature profiles. Such road signature profiles may be associated with any recognizable / measurable variation in at least one parameter associated with a road. For example, in some cases, such profiles may be associated with variations in road surface information, such as variations in surface roughness of a particular road segment, variations in road width across a particular road segment, variations in distances between dashed lines painted along a particular road segment, variations in road curvature along a particular road segment, etc. FIG. 11D shows an example of a road signature profile 1160. While the profile 1160 may represent any of the aforementioned parameters or others, in one example, the profile 1160 may represent a measure of road surface roughness, such as obtained by monitoring one or more sensors that provide outputs indicative of an amount of suspension displacement when a vehicle is driving a particular road segment.Alternatively or simultaneously, the profile 1160 may represent a variation in road width as determined based on image data obtained via a camera onboard a vehicle driving a particular road segment. Such profiles may be useful, for example, in determining a particular location of an autonomous vehicle with respect to a particular target trajectory. That is, when traversing a road segment, an autonomous vehicle may measure a profile associated with one or more parameters associated with the road segment. If the measured profile can be correlated / matched to a predetermined profile representing the parameter variation with respect to position along the road segment, the measured and predetermined profiles can be used (e.g., by overlaying corresponding portions of the measured and predetermined profiles) to determine a current position along the road segment and, therefore, a current position with respect to a target trajectory for the road segment.In some embodiments, sparse map 800 may include different trajectories based on different characteristics associated with a user of autonomous vehicles, environmental conditions, and / or other parameters related to the trip. For example, in some embodiments, different trajectories may be generated based on different user preferences and / or profiles. Sparse map 800 including such different trajectories may be provided to different autonomous vehicles of different users. For example, some users may prefer to avoid toll roads, while others prefer to take the shortest or fastest routes regardless of whether a toll road is present on the route. The disclosed systems may generate different sparse maps with different trajectories based on such different user preferences or profiles. As another example, some users may prefer to travel in a fast-driving lane, while others prefer to maintain a position in the central lane at all times.Different trajectories may be generated and enclosed in sparse map 800 based on different environmental conditions, such as day and night, snow, rain, fog, etc. Autonomous vehicles driving under different environmental conditions may be equipped with sparse map 800 generated based on these different environmental conditions. In some embodiments, cameras provided on autonomous vehicles may detect environmental conditions and may provide such information back to a server that generates and provides sparse maps. For example, the server may generate or update an already generated sparse map 800 to include trajectories that may be more suitable or safe for autonomous driving under the detected environmental conditions. The update of sparse map 800 based on environmental conditions may be performed dynamically as the autonomous vehicles are driving on roads.Other different parameters related to driving may also be used as a basis for generating and providing different sparse maps to different autonomous vehicles. For example, when an autonomous vehicle is traveling at a high speed, turns may be narrower. Trajectories associated with specific lanes, rather than roads, may be included in sparse map 800 such that the autonomous vehicle may maintain within a specific lane while the vehicle follows a specific trajectory. If an image captured by a camera onboard the autonomous vehicle indicates that the vehicle has drifted outside of the lane (e.g., crossed the lane marker), an action within the vehicle may be triggered to bring the vehicle back to the specified lane according to the specific trajectory.Crowdsourcing of a sparse cardThe disclosed sparse maps may be efficiently (and passively) generated by the force of crowdsourced. For example, any private or commercial vehicle equipped with a camera (e.g., a simple, low resolution camera, often included in present day vehicle first equipment) and a suitable image analysis processor may serve as the collection vehicle. No special equipment (e.g., high resolution imaging and / or positioning systems) is required. As a result of the disclosed crowdsourcing technique, the sparse maps generated may be highly accurate and include extremely refined position information (which enables navigation error limits of 10 cm or less) without requiring special imaging or scanning equipment as input to the map generation process. Crowdsourced operation also enables much faster (and less expensive) updates of the maps generated, as the mapping server system is constantly available new driving data from all roads traversed by private or commercial vehicles that are minimally equipped to also serve as collection vehicles. There is no need for special vehicles equipped with high resolution imaging and mapping sensors. Therefore, the costs associated with the construction of such special vehicles can be avoided. Further, updates to the sparse maps disclosed herein may be made much faster than systems relying on dedicated, specialized mapping vehicles (which, due to their cost and special equipment, are typically limited to a fleet of specialized vehicles that are many times less than the number of private or commercial vehicles already available for performing the disclosed collection techniques).The disclosed sparse maps generated by crowdsourced may be highly accurate because they are based on multiple inputs of several (10s, hundreds, millions, etc.) In addition, it is possible to generate collection vehicles that have collected driving data along a specific road segment. For example, each collection vehicle traveling along a particular road segment may record its actual trajectory and determine position information relative to detected objects / features along the road segment. This information is forwarded to a server by a plurality of collection vehicles. The actual trajectories are aggregated to generate a refined target trajectory for each valid travel path along the road segment. Additionally, the position information collected by the plurality of collection vehicles may also be aggregated for each of the detected objects / features along the road segment (semantic or non-semantic). As a result, the mapped position of each detected object / feature may average hundreds, thousands, or millions of individually determined positions for each detected object / feature. Such a technique can provide extremely accurate mapped positions for the detected objects / features.In some embodiments, the disclosed systems and methods may generate a sparse map for autonomous vehicle navigation. For example, disclosed systems and methods may use crowdsourced data to generate a sparseness that one or more autonomous vehicles may use to navigate along a system of roads. As used herein, "Crowdsourcing"means that data is received from different vehicles (e.g., autonomous vehicles) traveling at different times on a road segment, and such data is used to generate and / or update the road model, including sparse map tiles. The model or any of its sparse map tiles may in turn be transmitted to the vehicles or other vehicles that will travel later along the road segment to assist autonomous vehicle navigation. The road model may include a plurality of target trajectories that represent preferred trajectories that autonomous vehicles should follow as they traverse a road segment. The target trajectories may be the same as a reconstructed actual trajectory collected from a vehicle traversing a road segment that may be transmitted from the vehicle to a server. In some embodiments, the target trajectories may be different from actual trajectories that one or more vehicles previously taken when traversing a road segment. The target trajectories may be generated based on actual trajectories (e.g., by averaging or any other suitable operation).The vehicle trajectory data that a vehicle may upload to a server may correspond to the actual reconstructed trajectory for the vehicle or may correspond to a recommended trajectory that may be based on or refer to the actual reconstructed trajectory of the vehicle, but may be different from the actual reconstructed trajectory. For example, vehicles may modify their actual reconstructed trajectories and transmit (e.g., recommend) the modified actual trajectories to the server. The road model may include the recommended modified road models.Use trajectories as target trajectories for the autonomous navigation of other vehicles. In addition to trajectory information, other information for potential use in creating a sparse data map 800 may include information related to potential landmark candidates. For example, by crowdsourcing information, the disclosed systems and methods may identify potential landmarks in an environment and refine landmark positions. The landmarks may be used by an autonomous vehicle navigation system to determine and / or adjust the position of the vehicle along the target trajectories.The reconstructed trajectories that a vehicle can generate when the vehicle travels along a road can be obtained by any suitable method. In some embodiments, the reconstructed trajectories may be developed by merging motion segments for the vehicle using, e.g., ego motion estimation (e.g., three-dimensional translation and three-dimensional rotation of the camera and thus the body of the vehicle). The rotation and translation estimate may be determined based on the analysis of images captured by one or more image capture devices along with information from other sensors or devices, such as inertial sensors and speed sensors. For example, the inertial sensors may include an accelerometer or other suitable sensors configured to measure changes in translation and / or rotation of the vehicle body. The vehicle may include a speed sensor that measures a speed of the vehicle.In some embodiments, the own motion of the camera (and thus the vehicle body) may be estimated based on an optical flow analysis of the captured images. An optical flow analysis of an image sequence identifies the movement of pixels from the image sequence and determines the movements of the vehicle based on the identified movement. The own motion may be integrated over time and along the road segment to reconstruct a trajectory associated with the road segment followed by the vehicle.Data (e.g., reconstructed trajectories) collected from multiple vehicles in multiple trips along a road segment at different times may be used to construct the road model (e.g., including the target trajectories, etc.) included in sparse data map 800. Data collected from multiple vehicles in multiple trips along a road segment at different times may also be averaged to increase an accuracy of the model. In some embodiments, data regarding road geometry and / or landmarks may be received from multiple vehicles traveling through the common road segment at different times. Such data received from different vehicles may be combined to generate the road model and / or update the road model.The geometry of a reconstructed trajectory (and also a target trajectory) along a road segment may be represented by a three-dimensional space curve, which may be a spline connecting three-dimensional polynomials. The reconstructed trajectory curve may be determined from the analysis of a video stream or a plurality of images captured by a camera installed on the vehicle. In some embodiments, a location is identified in each frame or image that is a few meters ahead of the current position of the vehicle. This location is the location to which the vehicle is expected to travel in a predetermined period of time. This operation may be repeated frame by frame and, at the same time, the vehicle may calculate the own motion of the camera (rotation and translation). At each frame or image, a short range model for the desired path from the vehicle is generated in a frame of reference that is attached to the camera. The short range models may be merged to obtain a three-dimensional model of the road in a coordinate frame, which may be any or predetermined coordinate frame. The three-dimensional model of the road may then be adjusted by a spline, which may include or connect one or more polynomials of suitable orders.To complete the short-range road model at each frame, one or more recognition modules may be used. For example, a ground-up lane detection module may be used. The ground-up lane detection module may be useful when lane markings are dragged on the road. This module can search for edges in the image and join them together to form the lane markings. A second module may be used together with the ground-up lane detection module. The second module is an end-to-end deep neural network that can be trained to predict the correct short range path from an input image. In both modules, the road model can be recognized in the image coordinate system and converted into a three-dimensional space that can be virtually attached to the camera.Although the reconstructed trajectory modeling method may introduce a collection of errors due to the integration of self-motion over a long period of time, which may include a noise component, such errors may be insignificant as the generated model may provide sufficient accuracy for navigation over a local scale. In addition, it is possible to eliminate the integrated error by using external information sources such as satellite images or geodetic measurements. For example, the disclosed systems and methods may use a GNSS receiver to eliminate accumulated errors. However, the GNSS positioning signals may not always be available and accurate. The disclosed systems and methods may enable a steering application that is poorly dependent on the availability and accuracy of GNSS positioning. In such systems, the use of the GNSS signals may be limited. For example, in some embodiments, the disclosed systems may use the GNSS signals only for database indexing purposes.In some embodiments, the range scale (e.g., local scale) that may be relevant to autonomous vehicle navigation may be on the order of 50 meters, 100 meters, 200 meters, 300 meters, etc. Such distances may be used because the geometric road model is used primarily for two purposes: planning the forward trajectory and locating the vehicle on the road model. In some embodiments, the planning task may use the model over a typical range of 40 meters ahead (or another suitable distance ahead, such as 20 meters, 30 meters, 50 meters) when the control algorithm steers the vehicle according to a destination point that is 1.3 seconds ahead (or another time, such as 1.5 seconds, 1.7 seconds, 2 seconds, etc.). The localization task uses the road model over a typical range of 60 meters behind the car (or any other suitable distances, such as 50 meters, 100 meters, 150 meters, etc.) according to a method called "rear alignment," which is described in more detail in another section. The disclosed systems and methods may generate a geometric model that has sufficient accuracy over a particular range, such as 100 meters, such that a planned trajectory does not deviate from the lane center by more than, for example, 30 cm.As explained above, a three-dimensional road model can be constructed from detecting short-range sections and joining them together. The fusion may be enabled by computing a six-degree ego motion model using the video and / or images captured by the camera, data from the inertial sensors reflecting the movements of the vehicle, and the speed signal of the host vehicle. The accumulated error may be small enough over a local range scale, such as on the order of 100 meters. All this can be completed in a single trip over a particular road segment.In some embodiments, multiple trips may be used to average the resulting model and further increase its accuracy. The same car may travel the same route multiple times, or multiple cars may send their collected model data to a central server. In either case, a matching process may be performed to identify overlapping models and to allow averaging to generate target trajectories. The constructed model (e.g., including the target trajectories) may be used for steering once a convergence criterion is met. Subsequent trips may be used for further model improvements and to adapt to infrastructure changes.Sharing of driver experience (such as captured data) between multiple cars becomes feasible when connected to a central server. Each vehicle client may store a partial copy of a universal road model that may be relevant to its current location. A bidirectional update method between the vehicles and the server may be performed by the vehicles and the server. The small footprint concept discussed above enables the disclosed systems and methods to perform the bidirectional updates using very small bandwidth.Information relating to potential landmarks may also be determined and forwarded to a central server. For example, the disclosed systems and methods may determine one or more physical characteristics of a potential landmark based on one or more images including the landmark. The physical characteristics may include a physical size (e.g., height, width) of the landmark, a distance from a vehicle to a landmark, a distance between the landmark to a previous landmark, the lateral position of the landmark (e.g., the position of the landmark relative to the lane), the GPS coordinates of the landmark, a type of landmark, identification of text on the landmark, etc. For example, a vehicle may analyze one or more images captured by a camera to detect a potential landmark, such as a speed limit sign.The vehicle may determine a distance from the vehicle to the landmark or a position associated with the landmark (e.g., any semantic or non-semantic object or feature along a road segment) based on the analysis of the one or more images. In some embodiments, the distance may be determined based on the analysis of images of the landmark using a suitable image analysis method, such as a scaling method and / or an optical flow method. As noted above, a position of the object / feature may include a 2D image position (e.g., an X-Y pixel position in one or more captured images) of one or more points associated with the object / feature, or may include a real 3D position of one or more points (e.g., determined by structure-in-motion techniques / optical flow techniques, LIDAR or RADAR information, etc.). In some embodiments, the disclosed systems and methods may be configured to determine a type or classification of a potential landmark. If the vehicle determines that a particular potential landmark corresponds to a predetermined type or classification stored in a sparse map, it may be sufficient for the vehicle to communicate to the server an indication of the type or classification of the landmark along with its location. The server may store such information. At a later time, during navigation, a navigating vehicle may capture an image that includes a representation of the landmark, process the image (e.g., using a classifier), and compare the resulting landmark to confirm the detection of the mapped landmark, and use the mapped landmark in locating the navigating vehicle relative to the sparse map.In some embodiments, multiple autonomous vehicles traveling on a road segment may communicate with a server. The vehicles (or clients) may generate a curve describing their travel (e.g., by self-motion integration) in any coordinate frame. The vehicles can detect landmarks and locate them in the same frame. The vehicles may upload the turn and landmarks to the server. The server may collect data from vehicles about multiple trips and generate a uniform road model. For example, as discussed below with respect to FIG. 19, the server may generate a sparse map that includes the uniform road model using the uploaded curves and landmarks.The server may also distribute the model to clients (e.g., vehicles). For example, the server may distribute the sparse map to one or more vehicles. The server may update the model continuously or periodically as it receives new data from the vehicles. For example, the server may process the new data to evaluate whether the data includes information that should trigger an update or creation of new data on the server. The server may distribute the updated model or updates to the vehicles to provide autonomous vehicle navigation.The server may use one or more criteria to determine whether new data received from the vehicles should trigger an update to the model or trigger creation of new data. For example, if the new data indicates that a previously recognized landmark no longer exists at a particular location or is replaced with another landmark, the server may determine that the new data should trigger an update to the model. As another example, if the new data indicates that a road segment has been closed and if confirmed by data received from other vehicles, the server may determine that the new data should trigger an update to the model.The server may distribute the updated model (or portion of the model) to one or more vehicles traveling on the road segment with which the updates of the model are associated. The server may also distribute the updated model to vehicles about to travel on the road segment or vehicles whose planned trip includes the road segment with which the updates of the model are associated. For example, while an autonomous vehicle is traveling along another road segment before reaching the road segment with which an update is associated, the server may distribute the updates or the updated model to the autonomous vehicle before the vehicle reaches the road segment.In some embodiments, the remote server may collect trajectories and landmarks from multiple clients (e.g., vehicles traveling along a common road segment). The server may match curves using landmarks and generate an average road model based on the trajectories collected by the plurality of vehicles. The server may also calculate a graph of roads and the most likely path at each node or in connection with the road segment. For example, the remote server may align the trajectories to generate a sparse crowdsourced map from the collected trajectories.The server may average landmark characteristics received from multiple vehicles that have traveled along the common road segment, such as the distances between one landmark to another (e.g., a previous one along the road segment) as measured by multiple vehicles, to determine an arc length parameter and to assist location along the path and speed calibration for each client vehicle. The server may average the physical dimensions of a landmark measured by multiple vehicles traveling along the common road segment and recognize the same landmark. The averaged physical dimensions may be used to support distance estimation, such as the distance from the vehicle to the landmark. The server may average lateral positions of a landmark (e.g., position from the lane in which vehicles travel to the landmark) measured by multiple vehicles traveling along the common road segment and recognize the same landmark. The averaged lateral position may be used to assist lane assignment. The server may average the GPS coordinates of the landmark measured by multiple vehicles traveling along the same road segment and recognize the same landmark. The averaged GPS coordinates of the landmark may be used to assist in global localization or positioning of the landmark in the road model.In some embodiments, the server may identify model changes, such as constructions, diversions, new characters, removal of characters, etc., based on data received from the vehicles. The server may update the model continuously or periodically or immediately as it receives new data from the vehicles. The server may distribute updates to the model or the updated model to vehicles to provide autonomous navigation. For example, as discussed further below, the server may use crowdsourced data to filter out ghost landmarks detected by vehicles.In some embodiments, the server may analyze driver intervention during autonomous driving. The server may analyze data received from the vehicle at the time and location at which an intervention occurs and / or data received prior to the time at which the intervention occurred. The server may identify certain portions of the data that caused or closely related to the intervention, for example, data indicating a temporary means for closing the lane, data indicating a pedestrian on the road. The server may update the model based on the identified data. For example, the server may modify one or more trajectories stored in the model.FIG. 12 is a schematic illustration of a system that uses crowdsourcing to create a sparse map (as well as distribution and navigation using a sparse map created by crowdsourcing). FIG. 12 shows a road segment 1200 that includes one or more lanes. A plurality of vehicles 1205, 1210, 1215, 1220, and 1225 may travel on the road segment 1200 simultaneously or at different times (although shown in FIG. 12 as if they were to appear on the road segment 1200 simultaneously). At least one of vehicles 1205, 1210, 1215, 1220, and 1225 may be an autonomous vehicle. For simplicity of the present example, all vehicles 1205, 1210, 1215, 1220, and 1225 are assumed to be autonomous vehicles.Each vehicle may be similar to vehicles disclosed in other embodiments (e.g., vehicle 200), and may include components or devices included in or associated with vehicles disclosed in other embodiments. Each vehicle may be equipped with an image capture device or camera (e.g., image capture device 122 or camera 122). Each vehicle may communicate with a remote server 1230 over one or more networks (e.g., over a cellular network and / or the Internet, etc.) via wireless communication paths 1235, as indicated by the dashed lines. Each vehicle may transmit data to the server 1230 and receive data from the server 1230. For example, the server 1230 may collect data from multiple vehicles traveling at different times on the road segment 1200 and may process the collected data to generate a road navigation model of an autonomous vehicle or an update of the model. The server 1230 may transmit the road navigation model of the autonomous vehicle or the update of the model to the vehicles that have transmitted data to the server 1230. The server 1230 may transmit the autonomous vehicle road navigation model or the update of the model to other vehicles that are driving on the road segment 1200 at later times.When the vehicles 1205, 1210, 1215, 1220, and 1225 are traveling on the road segment 1200, navigation information collected (e.g., detected, acquired, or measured) by the vehicles 1205, 1210, 1215, 1220, and 1225 may be transmitted to the server 1230. In some embodiments, the navigation information may be associated with the common road segment 1200. The navigation information may include a trajectory associated with each of the vehicles 1205, 1210, 1215, 1220, and 1225 when each vehicle travels over the road segment 1200. In some embodiments, the trajectory may be reconstructed based on data captured by various sensors and devices provided on the vehicle 1205. For example, the trajectory may be reconstructed based on at least one of accelerometer data, speed data, landmark data, road geometry or profile data, vehicle positioning data, and ego motion data. In some embodiments, the trajectory may be reconstructed based on data from inertial sensors, such as accelerometers, and the speed of the vehicle 1205 detected by a speed sensor. Additionally, in some embodiments, the trajectory may be determined based on detected own motion of the camera (e.g., by a processor onboard each of the vehicles 1205, 1210, 1215, 1220, and 1225), which may indicate three-dimensional translation and / or three-dimensional rotations (or rotational motions). The own motion of the camera (and thus the vehicle body) may be determined from the analysis of one or more images captured by the camera.In some embodiments, the trajectory of the vehicle 1205 may be determined by a processor provided onboard the vehicle 1205 and transmitted to the server 1230. In other embodiments, the server 1230 may receive data captured by the various sensors and devices provided in the vehicle 1205 and determine the trajectory based on the data received from the vehicle 1205.In some embodiments, the navigation information transmitted from the vehicles 1205, 1210, 1215, 1220, and 1225 to the server 1230 may include data regarding the road surface, the road geometry, or the road profile. The geometry of the road segment 1200 may include a lane structure and / or landmarks. The structure of the lane may include the total number of lanes of the road segment 1200, the type of lanes (e.g., one-way road, oncoming road, lane, passing lane, etc.), markings on the lanes, the width of the lanes, etc. In some embodiments, the navigation information may include a lane assignment, e.g., which lane of a plurality of lanes a vehicle is traveling. For example, the lane assignment may be assigned a numerical value "3" indicating that the vehicle is traveling in the third lane from the left or right. As another example, the lane assignment may be associated with a text value "center lane" indicating that the vehicle is driving in the center lane.The server 1230 may store the navigation information on a non-transitory computer readable medium, such as a hard disk, a compact disc, a tape, a memory, etc. The server 1230 may generate (e.g., by a processor included in the server 1230) at least a portion of an autonomous vehicle road navigation model for the common road segment 1200 based on the navigation information received from the plurality of vehicles 1205, 1210, 1215, 1220, and 1225, and may store the model as a portion of a sparse map. The server 1230 may determine a trajectory associated with each lane based on crowdsourced data (e.g., navigation information) received from multiple vehicles (e.g., 1205, 1210, 1215, 1220, and 1225) traveling on a lane of the road segment at different times. The server 1230 may generate the autonomous vehicle road navigation model or a portion of the model (e.g., an updated portion) based on a plurality of trajectories determined based on the crowdsourcing navigation data. The server 1230 may transmit the model or updated portion of the model to one or more autonomous vehicles 1205, 1210, 1215, 1220, and 1225 traveling on the road segment 1200 or any other autonomous vehicles traveling on the road segment at a later time to update an existing autonomous vehicle road navigation model provided in a navigation system of the vehicles. The autonomous vehicle road navigation model may be used by the autonomous vehicles in autonomously navigating along the common road segment 1200.As discussed above, the autonomous vehicle road navigation model may be included in a sparse map (e.g., sparse map 800 shown in FIG. 8). Sparse map 800 may include a sparse record of data related to road geometry and / or landmarks along a road that may provide sufficient information to guide autonomous navigation of an autonomous vehicle, but do not require excessive data storage. In some embodiments, the autonomous vehicle road navigation model may be stored separately from sparse map 800 and may use map data from sparse map 800 when the model is executed for navigation. In some embodiments, the autonomous vehicle road navigation model may use map data included in sparse map 800 to determine target trajectories along road segment 1200 to guide autonomous navigation of autonomous vehicles 1205, 1210, 1215, 1220, and 1225, or other vehicles that will later travel along road segment 1200. For example, if the autonomous vehicle road navigation model is executed by a processor included in a vehicle 1205 navigation system, the model may cause the processor to compare the trajectories determined based on the navigation information received from the vehicle 1205 with predetermined trajectories included in the sparse map 800 to validate and / or correct the current driving history of the vehicle 1205.In the autonomous vehicle road navigation model, the geometry of a road feature or target trajectory may be encoded by a curve in a three-dimensional space. In one embodiment, the curve may be a three-dimensional spline that includes one or more connecting three-dimensional polynomials. As one of ordinary skill in the art would understand, a spline may be a numerical function piecewise defined by a series of polynomials for fitting data. A spline for adjusting the three-dimensional geometry data of the road may include a linear spline (first order), a quadratic spline (second order), a cubic spline (third order), or any other splines (other orders), or a combination thereof. The spline may include one or more three-dimensional polynomials of different orders that connect (e.g., fit) data points of the three-dimensional geometry data of the road. In some embodiments, the autonomous vehicle road navigation model may include a three-dimensional spline corresponding to a target trajectory along a common road segment (e.g., road segment 1200) or a lane of travel of the road segment 1200.As discussed above, the autonomous vehicle road navigation model included in the sparse map may include other information, such as identification of at least one landmark along the road segment 1200. The landmark may be visible in a field of view of a camera (e.g., camera 122) installed on each of the vehicles 1205, 1210, 1215, 1220, and 1225. In some embodiments, camera 122 may capture an image of a landmark. A processor (e.g., processor 180, 190 or processing unit 110) provided on the vehicle 1205 may process the image of the landmark to extract identification information for the landmark. The landmark identification information may be stored in the sparse map 800 instead of an actual image of the landmark. The landmark identification information may require much less storage than an actual image. Other sensors or systems (e.g., GPS system) may also provide certain landmark identification information (e.g., landmark position). The landmark may include at least one of a road sign, an arrow mark, a lane mark, a dashed lane mark, a traffic light, a stop line, a directional sign (e.g., a highway exit sign with an arrow indicating a direction, a highway sign with arrows pointing in different directions or locations), a landmark beacon, or a lamp post. A landmark beacon refers to a device (e.g., an RFID device) installed along a road segment that transmits or reflects a signal to a receiver installed on a vehicle such that when the vehicle passes the device, the beacon received from the vehicle and the location of the device (e.g., determined from the GPS location of the device) may be used as a landmark to be included in the autonomous vehicle road navigation model and / or sparse map 800.The identification of at least one landmark may include a position of the at least one landmark. The position of the landmark may be determined based on position measurements made using sensor systems (e.g., global positioning systems, inertial positioning systems, landmark beacon, etc.) associated with the plurality of vehicles 1205, 1210, 1215, 1220, and 1225. In some embodiments, the position of the landmark may be determined by averaging the position measurements detected, collected, or received by sensor systems on different vehicles 1205, 1210, 1215, 1220, and 1225 through multiple trips. For example, the vehicles 1205, 1210, 1215, 1220, and 1225 may transmit position measurement data to the server 1230, which may average the position measurements and use the averaged position measurement as the position of the landmark. The position of the landmark may be continuously refined by measurements received from vehicles in subsequent trips.The identification of the landmark may include a size of the landmark. The processor provided on a vehicle (e.g., 1205) may estimate the physical size of the landmark based on the analysis of the images. The server 1230 may receive multiple estimates of the physical size of the same landmark from different vehicles over different trips. The server 1230 may average the different estimates to arrive at a physical quantity for the landmark and store that landmark quantity in the road model. The estimation of the physical quantity may be used to further determine or estimate a distance from the vehicle to the landmark. The distance to the landmark may be estimated based on the current speed of the vehicle and an expansion scale based on the position of the landmark appearing in the images with respect to the expansion focus of the camera. For example, the distance to the landmark may be estimated by Z=V*dt*R / D, where V is the speed of the vehicle, R is the distance in the image from the landmark at time t 1 to the focus of expansion, and D represents the change in the distance for the landmark in the image from t 1 to t 2. dt (t 2-t 1). For example, the distance to the landmark may be estimated by Z=V*dt*R / D, where V is the speed of the vehicle, R is the distance in the image between the landmark and the extension focus, dt is a time interval, and D is the image shift of the landmark along the epipolar line. Other equations corresponding to the above equation, such as Z=V*ω / Δω, may be used to estimate the distance to the landmark. Here, V is the vehicle speed, ω is an image length (such as the object width), and Δω is the change in this image length in a unit time.If the physical size of the landmark is known, the distance to the landmark may also be determined based on the following equation: Z=f*W / ω, where f is the focal length, W is the size of the landmark (e.g., height or width), ω is the number of pixels when the landmark exits the image. From the above equation, a change in distance Z can be calculated using ΔZ=f*W*Δω / ω2+f*ΔW / ω, where ΔW drops to zero by averaging, and where Δω is the number of pixels representing bounding box accuracy in the image. A value that estimates the physical size of the landmark may be calculated by averaging multiple observations on the server side. The resulting error in the range estimation may be very small. There are two sources of error that may occur when using the above formula, namely ΔW and Δω. Their contribution to the range error is given by ΔZ=f*W*Δω / ω2+f*ΔW / ω. However, ΔW drops to zero by averaging; therefore, ΔZ is determined by Δω (e.g., the imprecision of the bounding box in the image).For landmarks of unknown dimensions, the distance to the landmark can be estimated by tracking feature points on the landmark between successive frames. For example, certain features appearing on a speed limit sign may be tracked between two or more frames. Based on these tracked features, a distance distribution per feature point may be generated. The distance estimate may be extracted from the distance distribution. For example, the most common distance that appears in the distance distribution may be used as the distance estimate. As another example, the average of the distance distribution may be used as the distance estimate.FIG. 13 illustrates an example of an autonomous vehicle road navigation model represented by a plurality of three-dimensional splines 1301, 1302, and 1303. The curves 1301, 1302, and 1303 shown in FIG. 13 are for illustrative purposes only. Each spline may include one or more three-dimensional polynomials that connect a plurality of data points 1310. Each polynomial may be a first order polynomial, a second order polynomial, a third order polynomial, or a combination of any suitable polynomials having different orders. Each data point 1310 may be associated with the navigation information received from the vehicles 1205, 1210, 1215, 1220, and 1225. In some embodiments, each data point 1310 may be associated with data relating to landmarks (e.g., size, location, and identification information of landmarks) and / or road signature profiles (e.g., road geometry, road roughness profile, road curvature profile, road width profile). In some embodiments, some data points 1310 may be associated with data relating to landmarks and others may be associated with data relating to road signature profiles.FIG. 14 illustrates the raw data to location 1410 (e.g., GPS data) received from five different drives. A trip may be separate from another trip if it was simultaneously crossed by separate vehicles, at separate times from the same vehicle, or at separate times from separate vehicles. To account for errors in the location data 1410 and for different locations of vehicles within the same lane (e.g., one vehicle may travel closer to the left side of a lane than another), the server 1230 may generate a map skeleton 1420 using one or more statistical techniques to determine whether variations in the raw location data 1410 represent actual deviations or statistical errors. Each path within the skeleton 1420 may be linked back to the raw data 1410 that formed the path. For example, the path between A and B within the skeleton 1420 is associated with the raw data 1410 from the trips 2, 3, 4, and 5, but not from the trip 1. The skeleton 1420 may not be detailed enough to be used to navigate a vehicle (e.g., because it combines multiple lane drives on the same road as opposed to the splines described above), but may provide useful topological information and may be used to define intersections.FIG. 15 illustrates an example by which additional detail may be generated for a sparse map within a segment of a map skeleton (e.g., segments A-B within skeleton 1420). As shown in FIG. 15, the data (e.g., ego motion data, road marker data, etc.) may be shown as a function of position S (or S 1 or S 2) along the trip. The server 1230 may identify landmarks for the sparse map by identifying unique matches between the landmarks 1501, 1503, and 1505 of the trip 1510 and the landmarks 1507 and 1509 of the trip 1520. Such a matching algorithm may result in the identification of landmarks 1511, 1513, and 1515. However, one skilled in the art would recognize that other matching algorithms may be used. For example, probability optimization may be used instead of or in combination with a unique match. The server 1230 may align the trips longitudinally to align the aligned landmarks. For example, the server 1230 may select one trip (e.g., trip 1520) as a reference trip and then shift the other trip(s) (e.g., trip 1510) for alignment and / or elastically expand.Figure 16 shows an example of aligned landmark data for use in a sparse map. In the example of FIG. 16, landmark 1610 includes a road sign. The example in FIG. 16 further shows data of a plurality of trips 1601, 1603, 1605, 1607, 1609, 1611, and 1613. In the example of FIG. 16, the data from the trip 1613 consists of a "ghost" landmark, and the server 1230 may identify it as such, as none of the trips 1601, 1603, 1605, 1607, 1609, and 1611 include identification of a landmark near the identified landmark in the trip 1613. Accordingly, the server 1230 may accept potential landmarks when a ratio of images in which the landmark appears to images in which the landmark does not appear exceeds a threshold, and / or reject potential landmarks when a ratio of images in which the landmark does not appear to images in which the landmark appears exceeds a threshold.FIG. 17 illustrates a travel data generation system 1700 that may be used to crowdsource a sparse map. As shown in FIG. 17, system 1700 may include a camera 1701 and a device 1703 for locating (e.g., a GPS locator). The camera 1701 and the locator 1703 may be mounted on a vehicle (e.g., one of the vehicles 1205, 1210, 1215, 1220, and 1225). The camera 1701 may generate a plurality of data of a plurality of types, for example, own motion data, traffic sign data, road data, or the like. The camera data and location data may be segmented into trip segments 1705. For example, the travel segments 1705 may each have camera data and location data of less than 1 km travel.In some embodiments, system 1700 may eliminate redundancies in travel segments 1705. For example, if an landmark appears in multiple images from camera 1701, system 1700 may remove the redundant data such that travel segments 1705 include only a copy of the location and any metadata related to the landmark. As another example, if a lane marker appears in multiple images from camera 1701, system 1700 may remove the redundant data such that travel segments 1705 include only one copy of the location and any metadata related to the lane marker.The system 1700 also includes a server (e.g., server 1230). The server 1230 may receive travel segments 1705 from the vehicle and recombine the travel segments 1705 into a single travel 1707. Such an arrangement may allow for reduced bandwidth requirements in transferring data between the vehicle and the server, while also allowing the server to store data relating to an entire trip.FIG. 18 illustrates the system 1700 of FIG. 17, further configured for crowdsourced map browsing. As in FIG. 17, system 1700 includes vehicle 1810 capturing trip data using, for example, a camera (e.g., producing ego motion data, road sign data, road data, or the like) and a locator (e.g., a GPS locator). As in FIG. 17, vehicle 1810 divides the collected data into travel segments (shown in FIG. 18 as "DS1 1", "DS2 1", "DSN 1"). Server 1230 then receives the trip segments and reconstructs a trip from the received segments (shown as "trip 1" in Figure 18).As further shown in FIG. 18, system 1700 also receives data from other vehicles. For example, the vehicle 1820 also captures travel data using, for example, a camera (e.g., generating ego motion data, road sign data, road data, or the like) and a locator (e.g., a GPS locator). Similar to the vehicle 1810, the vehicle 1820 segments the collected data into trip segments (shown as "DS1 2", "DS22", "DSN 2" in FIG. 18 ). Server 1230 then receives the trip segments and reconstructs a trip from the received segments (shown as "trip 2" in Figure 18). Any number of additional vehicles may be used. For example, FIG. 18 also includes "CAR N" that captures trip data, divides it into trip data segments (shown as "DS1 N", "DS2N", "DSN N" in FIG. 18 ), and sends it to server 1230 to reconstruct it into a trip (shown as "Drive N" in FIG. 18).As shown in FIG. 18, server 1230 may create a sparse map (shown as "MAP") using the reconstructed trips (e.g., "trip 1", "trip 2", and "trip N") collected from a plurality of vehicles (e.g., "AUTO 1" (also referred to as vehicle 1810), "AUTO 2" (also referred to as vehicle 1820), and "AUTO N").FIG. 19 is a flow diagram illustrating an example process 1900 for generating a sparse map for autonomous vehicle navigation along a road segment. The process 1900 may be performed by one or more processing devices included in the server 1230.The process 1900 may include receiving a plurality of images captured when one or more vehicles cross the road segment (step 1905). The server 1230 may receive images from cameras included in one or more of the vehicles 1205, 1210, 1215, 1220, and 1225. For example, the camera 122 may capture one or more images of the environment surrounding the vehicle 1205 as the vehicle 1205 travels along the road segment 1200. In some embodiments, server 1230 may also receive reduced image data having redundancies removed by a processor on vehicle 1205, as discussed above with respect to FIG. 17.The process 1900 may further include identifying, based on the plurality of images, at least one line representation of a feature of the road surface extending along the road segment (step 1910). Each line representation may represent a path along the road segment that substantially corresponds to the road surface feature. For example, the server 1230 may analyze the environmental images received from the camera 122 to identify a road edge or lane marker and determine a travel trajectory along the road segment 1200 associated with the road edge or lane marker. In some embodiments, the trajectory (or line representation) may include a spline, polynomial representation, or curve. The server 1230 may determine the travel trajectory of the vehicle 1205 based on camera own movements (e.g., three-dimensional translation and / or three-dimensional rotational movements) received at step 1905.The process 1900 may also include identifying, based on the plurality of images, a plurality of landmarks associated with the road segment (step 1910). For example, the server 1230 may analyze the environmental images received from the camera 122 to identify one or more landmarks, such as the road sign along the road segment 1200. The server 1230 may identify the landmarks using the analysis of the plurality of images captured when one or more vehicles cross the road segment. To enable crowdsourcing, the analysis may include rules regarding acceptance and rejection of possible landmarks associated with the road segment. For example, the analysis may include accepting potential landmarks when a ratio of images in which the landmark appears to images in which the landmark does not appear exceeds a threshold, and / or rejecting potential landmarks when a ratio of images in which the landmark does not appear to images in which the landmark appears exceeds a threshold.The process 1900 may include other operations or steps performed by the server 1230. For example, the navigation information may include a target trajectory for vehicles to travel along a road segment, and the process 1900 may include clustering, by the server 1230, vehicle trajectories related to multiple vehicles traveling on the road segment and determining the target trajectory based on the clustered vehicle trajectories, as discussed in more detail below. Clustering vehicle trajectories may include clustering, by the server 1230, the plurality of trajectories related to the vehicles traveling on the road segment into a plurality of clusters based on at least one of the absolute traveling direction of the vehicles or the lane assignment of the vehicles. Generating the target trajectory may include averaging, by the server 1230, the clustered trajectories. As another example, process 1900 may include aligning data received at step 1905. Other processes or steps performed by the server 1230 as described above may also be included in process 1900.The disclosed systems and methods may include other features. For example, the disclosed systems may use local coordinates instead of global coordinates. For autonomous driving, some systems may present data in world coordinates. For example, length and width coordinates on the surface of the earth may be used. To use the map for steering, the host vehicle may determine its position and orientation relative to the map. It will of course appear to use an on-board GPS device to position the vehicle on the map and to find the rotational transformation between the body reference frame and the world reference frame (e.g., north, east, and down). Once the body reference frame is aligned with the map reference frame, the desired route in the body reference frame may be expressed and the steering commands may be calculated or generated.The disclosed systems and methods may enable autonomous vehicle navigation (e.g., steering control) with low footprint models that may be collected by the autonomous vehicles themselves without the aid of expensive survey equipment. To assist autonomous navigation (e.g., steering applications), the road model may include a sparse map with the geometry of the road, its lane structure, and landmarks that may be used to determine the location or position of vehicles along a trajectory included in the model. As discussed above, the generation of the sparse map may be performed by a remote server that communicates with vehicles traveling on the road that receive data from the vehicles. The data may include collected data, trajectories reconstructed based on the collected data, and / or recommended trajectories that may represent modified reconstructed trajectories. As discussed below, the server may transmit the model back to the vehicles or other vehicles that will later travel on the road to assist autonomous navigation.FIG. 20 illustrates a block diagram of the server 1230. The server 1230 may include a communication unit 2005, which may include both hardware components (e.g., communication control circuits, switches, and antennas) and software components (e.g., communication protocols, computer codes). For example, the communication unit 2005 may include at least one network interface. The server 1230 may communicate with the vehicles 1205, 1210, 1215, 1220, and 1225 via the communication unit 2005. For example, the server 1230 may receive navigation information transmitted from the vehicles 1205, 1210, 1215, 1220, and 1225, via the communication unit 2005. The server 1230 may distribute the road navigation model of the autonomous vehicle to one or more autonomous vehicles via the communication unit 2005.The server 1230 may include at least one non-volatile storage medium 2010, such as a hard disk, a compact disc, a tape, etc. The storage device 1410 may be configured to store data such as navigation information received from the vehicles 1205, 1210, 1215, 1220, and 1225, and / or the autonomous vehicle road navigation model that the server 1230 generates based on the navigation information. The storage device 2010 may be configured to store any other information, such as a sparse map (e.g., sparse map 800 discussed above with respect to FIG. 8).In addition to or in place of storage device 2010, server 1230 may include a memory 2015. The memory 2015 may be similar to or different from the memory 140 or 150. The memory 2015 may be nonvolatile memory such as flash memory, random access memory, etc. The memory 2015 may be configured to store data such as computer code or instructions executable by a processor (e.g., the processor 2020), map data (e.g., sparse map data 800), the autonomous vehicle road navigation model, and / or navigation information received from the vehicles 1205, 1210, 1215, 1220, and 1225.The server 1230 may include at least one processing device 2020 configured to execute computer code or instructions stored in the memory 2015 to perform various functions. For example, the processing device 2020 may analyze the navigation information received from the vehicles 1205, 1210, 1215, 1220, and 1225, and generate the road navigation model of the autonomous vehicle based on the analysis. The processing device 2020 may control the communication unit 1405 to distribute the road navigation model of the autonomous vehicle to one or more autonomous vehicles (e.g., one or more of the vehicles 1205, 1210, 1215, 1220, and 1225, or any vehicle that travels on the road segment 1200 at a later time). The processing device 2020 may be similar to or different from the processor 180, 190 or the processing unit 110.FIG. 21 illustrates a block diagram of the memory 2015, which may store computer code or instructions for performing one or more operations to generate a road navigation model for use in autonomous vehicle navigation. As shown in FIG. 21, the memory 2015 may store one or more modules to perform the processes for processing car navigation information. For example, the memory 2015 may include a model generation module 2105 and a model distribution module 2110. The processor 2020 may execute the instructions stored in any of the modules 2105 and 2110 included in the memory 2015.The model generation module 2105 may store instructions that, when executed by the processor 2020, may generate at least a portion of an autonomous vehicle road navigation model for a common road segment (e.g., the road segment 1200) based on navigation information received from the vehicles 1205, 1210, 1215, 1220, and 1225. For example, in generating the autonomous vehicle road navigation model, the processor 2020 may cluster vehicle trajectories into different clusters along the common road segment 1200. The processor 2020 may determine a target trajectory along the common road segment 1200 based on the clustered vehicle trajectories for each of the various clusters. Such an operation may include finding a mean or average trajectory of the clustered vehicle trajectories (e.g., by averaging data representing the clustered vehicle trajectories) in each cluster. In some embodiments, the target trajectory may be associated with a single lane of the common road segment 1200.The road model and / or sparse map may store trajectories associated with a road segment. These trajectories may be referred to as target trajectories provided to autonomous vehicles for autonomous navigation. The target trajectories may be received from multiple vehicles or may be generated based on actual trajectories or recommended trajectories (actual trajectories with some modifications) received from multiple vehicles. The target trajectories included in the road model or sparse map may be continuously updated (e.g., averaged) with new trajectories received from other vehicles.Vehicles traveling on a road segment may collect data by various sensors. The data may include landmarks, road signature profile, vehicle motion (e.g., accelerometer data, speed data), vehicle position (e.g., GPS data), and may either reconstruct the actual trajectories themselves or transmit the data to a server that reconstructs the actual trajectories for the vehicles. In some embodiments, the vehicles may transmit data related to a trajectory (e.g., a curve in any reference frame), landmark data, and lane assignment along the travel path to the server 1230. Different vehicles traveling along the same road segment during multiple trips may have different trajectories. The server 1230 may identify routes or trajectories associated with each lane from the trajectories received from vehicles through a clustering process.FIG. 22 illustrates a process for clustering vehicle trajectories associated with vehicles 1205, 1210, 1215, 1220, and 1225 to determine a target trajectory for the common road segment (e.g., road segment 1200). The target trajectory or a plurality of target trajectories determined from the clustering process may be included in the autonomous vehicle road navigation model or sparse map 800. In some embodiments, vehicles 1205, 1210, 1215, 1220, and 1225 traveling along road segment 1200 may transmit a plurality of trajectories 2200 to server 1230. In some embodiments, server 1230 may generate trajectories based on landmark, road geometry, and vehicle motion information received from vehicles 1205, 1210, 1215, 1220, and 1225. To generate the autonomous vehicle road navigation model, the server 1230 may cluster vehicle trajectories 1600 into a plurality of clusters 2205, 2210, 2215, 2220, 2225, and 2230, as shown in FIG. 22.Clustering may be performed using various criteria. In some embodiments, all trips in a cluster may be similar with respect to the absolute direction of travel along road segment 1200. The absolute heading direction may be obtained from GPS signals received from vehicles 1205, 1210, 1215, 1220, and 1225. In some embodiments, the absolute heading direction may be obtained using dead reckoning navigation. Dead reckoning navigation, as would be understood by one of ordinary skill in the art, may be used to determine the current position, and thus the heading direction, of the vehicles 1205, 1210, 1215, 1220, and 1225, using previously determined position, estimated speed, etc. Trajectories clustered by the absolute heading may be useful for identifying routes along the roads.In some embodiments, all trips in a cluster may be similar in terms of lane assignment (e.g., in the same lane before and after an intersection) along the trip on road segment 1200. Trajectories clustered by the lane assignment may be useful for identifying lanes along the lanes. In some embodiments, both criteria (e.g., absolute heading and lane assignment) may be used for clustering.In each cluster 2205, 2210, 2215, 2220, 2225, and 2230, trajectories may be averaged to obtain a target trajectory associated with the specific cluster. For example, the trajectories of multiple trips associated with the same lane cluster may be averaged. The averaged trajectory may be a target trajectory associated with a specific lane. To average a cluster of trajectories, the server 1230 may select a reference frame of any trajectory C 0. For all other trajectories (C1,..., Cn), the server 1230 may find a rigid transformation mapping Ci to C0, where i=1, 2,..., n, n being a positive integer corresponding to the total number of trajectories included in the cluster. The server 1230 may calculate a mean curve or trajectory in the C0 reference frame.In some embodiments, the landmarks may define an arc length alignment between different trips that may be used to align trajectories with lanes. In some embodiments, lane markings before and after an intersection may be used to align trajectories with lanes.To compile lanes from the trajectories, the server 1230 may select a reference frame of any lane. The server 1230 may map partially overlapping lanes to the selected reference frame. The server 1230 may continue mapping until all lanes are in the same reference frame. Lanes that are adjacent to each other may be aligned as if they were the same lane, and later they may be laterally shifted.Landmarks detected along the road segment may be mapped to the common reference frame, first on the lane plane, then on the intersection plane. For example, the same landmarks may be detected multiple times by multiple vehicles in multiple trips. The data regarding the same landmarks received in different trips may be slightly different. Such data may be averaged and mapped to the same reference frame, such as the C0 reference frame. Additionally or alternatively, the variance of the same landmark data received in multiple trips may be calculated.In some embodiments, each lane of the road segment 120 may be associated with a target trajectory and certain landmarks. The target trajectory or a plurality of such target trajectories may be included in the autonomous vehicle's road navigation model, which may be used later by other autonomous vehicles traveling along the same road segment 1200. Landmarks identified by the vehicles 1205, 1210, 1215, 1220, and 1225 as the vehicles travel along the road segment 1200 may be recorded in association with the target trajectory.The data of the target trajectories and landmarks may be continuously or periodically updated with new data received from other vehicles in subsequent trips. To locate an autonomous vehicle, the disclosed systems and methods may use an extended Kalman filter. The location of the vehicle can be determined based on three-dimensional position data and / or three-dimensional orientation data, prediction of the future location in front of the current location of the vehicle by integration of the own movement. The location of the vehicle can be corrected or adjusted by image observations of landmarks. For example, if the vehicle detects an landmark within an image captured by the camera, the landmark may be compared to a known landmark stored within the road model or sparse map 800. The known landmark may include a known location (e.g., GPS data) along a target trajectory stored in the road model and / or sparse map 800. Based on the current speed and the images of the landmark, the distance from the vehicle to the landmark may be estimated. The location of the vehicle along a target trajectory may be adjusted based on the distance to the landmark and the known location of the landmark (stored in the road model or sparse map 800). It can be assumed that the position / location data of the landmark (e.g., averages of multiple trips) stored in the road model and / or sparse map 800 is accurate.In some embodiments, the disclosed system may form a closed loop subsystem in which the estimation of the six degree-of-freedom location of the vehicle (e.g., three-dimensional position data plus three-dimensional orientation data) may be used to navigate (e.g., steer the wheel of) the autonomous vehicle to reach a desired point (e.g., 1.3 seconds before in the stored). In turn, data measured from steering and actual navigation may be used to estimate the location with six degrees of freedom.In some embodiments, poles along a road, such as lamp posts and power or cable line poles, may be used as landmarks for locating the vehicles. Other landmarks, such as traffic signs, traffic lights, arrows on the road, stop lines, and static features or signatures of an object along the road segment may also be used as landmarks for locating the vehicle. When poles are used for localization, the x-observation of the poles (i.e., the angle of view from the vehicle) may be used instead of the y-observation (i.e., the distance from the pole) because the floors of the poles may be hidden and sometimes do not lie on the road plane.FIG. 23 illustrates a navigation system for a vehicle that may be used for autonomous navigation using a sparse crowdsouring map. For purposes of illustration, the vehicle is referenced as vehicle 1205. The vehicle shown in FIG. 23 may be any other vehicle disclosed herein, including, for example, vehicles 1210, 1215, 1220, and 1225, as well as vehicle 200 shown in other embodiments. As shown in FIG. 12, vehicle 1205 may communicate with server 1230. The vehicle 1205 may include an image capture device 122 (e.g., camera 122). The vehicle 1205 may include a navigation system 2300 configured to provide navigation guidance to the vehicle 1205 to travel on a road (e.g., road segment 1200). The vehicle 1205 may also include other sensors, such as a speed sensor 2320 and an accelerometer 2325. The speed sensor 2320 may be configured to detect the speed of the vehicle 1205. The accelerometer 2325 may be configured to detect acceleration or deceleration of the vehicle 1205. The vehicle 1205 shown in FIG. 23 may be an autonomous vehicle, and the navigation system 2300 may be used to provide navigation guidance for autonomous driving. Alternatively, the vehicle 1205 may also be a non-autonomous human-controlled vehicle and the navigation system 2300 may further be used to provide navigation guidance.The navigation system 2300 may include a communication unit 2305 configured to communicate with the server 1230 via the communication path 1235. The navigation system 2300 may also include a GPS unit 2310 configured to receive and process GPS signals. The navigation system 2300 may further include at least one processor 2315 configured to process data such as GPS signals, map data from sparse map 800 (which may be stored on a storage device provided onboard vehicle 1205 and / or received from server 1230), road geometry captured by a road profile sensor 2330, images captured by camera 122, and / or an autonomous vehicle road navigation model received from server 1230. The road profile sensor 2330 may include various types of devices for measuring various types of road profiles, such as roughness of the road surface, road width, bump of the road, road curvature, etc. For example, the road profile sensor 2330 may include a device that measures the motion of a suspension of a vehicle 2305 to derive the road unevenness profile. In some embodiments, the road profile sensor 2330 may include radar sensors to measure the distance from the vehicle 1205 to the sides of the road (e.g., barrier on the sides of the road), thereby measuring the width of the road. In some embodiments, the road profile sensor 2330 may include a device configured to measure the height of the road up and down. In some embodiments, the road profile sensor 2330 may include a device configured to measure road curvature. For example, a camera (e.g., camera 122 or other camera) may be used to capture images of the road showing road curvatures. The vehicle 1205 may use such images to detect road curvatures.The at least one processor 2315 may be programmed to receive from the camera 122 at least one environmental image associated with the vehicle 1205. The at least one processor 2315 may analyze the at least one environmental image to determine navigation information related to the vehicle 1205. The navigation information may include a trajectory related to the travel of the vehicle 1205 along the road segment 1200. The at least one processor 2315 may determine the trajectory based on movements of the camera 122 (and thus the vehicle), such as three-dimensional translation and three-dimensional rotational movements. In some embodiments, the at least one processor 2315 may determine the translation and rotational movements of the camera 122 based on the analysis of a plurality of images captured by the camera 122. In some embodiments, the navigation information may include lane assignment information (e.g., which lane the vehicle 1205 is traveling along the road segment 1200). The navigation information transmitted from the vehicle 1205 to the server 1230 may be used by the server 1230 to generate and / or update a road navigation model of the autonomous vehicle, which may be transmitted back from the server 1230 to the vehicle 1205 to provide autonomous navigation guidance for the vehicle 1205.The at least one processor 2315 may also be programmed to transmit the navigation information from the vehicle 1205 to the server 1230. In some embodiments, the navigation information may be transmitted to the server 1230 along with the road information. The road location information may include at least one of GPS signals received from the GPS device 2310, landmarks information, road geometry, lanes, etc. The at least one processor 2315 may receive from the server 1230 the autonomous vehicle road navigation model or a portion of the model. The autonomous vehicle road navigation model received from the server 1230 may include at least one update based on the navigation information transmitted from the vehicle 1205 to the server 1230. The portion of the model transmitted from the server 1230 to the vehicle 1205 may include an updated portion of the model. The at least one processor 2315 may cause at least one navigation maneuver (e.g., steer, such as turning, braking, accelerating, passing another vehicle, etc.) by the vehicle 1205 based on the received autonomous vehicle road navigation model or the updated portion of the model.The at least one processor 2315 may be configured to communicate with various sensors and components included in the vehicle 1205, including the communication unit 2305, the GPS unit 2310, the camera 122, the speed sensor 2320, the accelerometer 2325, and the road profile sensor 2330. The at least one processor 2315 may collect information or data from various sensors and components and transmit the information or data to the server 1230 via the communication unit 2305. Alternatively or additionally, various sensors or components of the vehicle 1205 may also communicate with the server 1230 and transmit data or information collected by the sensors or components to the server 1230.In some embodiments, vehicles 1205, 1210, 1215, 1220, and 1225 may communicate with each other and may share navigation information with each other, such that at least one of vehicles 1205, 1210, 1215, 1220, and 1225 may generate the autonomous vehicle road navigation model using crowdsourced, e.g., based on information shared by other vehicles. In some embodiments, vehicles 1205, 1210, 1215, 1220, and 1225 may share navigation information with each other, and each vehicle may update its own road navigation model of the autonomous vehicle provided in the vehicle. In some embodiments, at least one of vehicles 1205, 1210, 1215, 1220, and 1225 (e.g., vehicle 1205) may function as a lift vehicle. The at least one processor 2315 of the hub vehicle (e.g., the vehicle 1205) may perform some or all of the functions performed by the server 1230. For example, the at least one processor 2315 of the hub vehicle may communicate with other vehicles and receive navigation information from other vehicles. The at least one processor 2315 of the hub vehicle may generate the autonomous vehicle road navigation model or an update of the model based on the shared information received from other vehicles. The at least one processor 2315 of the hub vehicle may transmit the autonomous vehicle road navigation model or the update of the model to other vehicles to provide autonomous navigation guidance.Navigation Based on Sparse MapsAs discussed above, the autonomous vehicle road navigation model that includes sparse map 800 may include a plurality of mapped lane markings and a plurality of mapped objects / features associated with a road segment. As discussed in more detail below, these mapped lane markings, objects, and features may be used when the autonomous vehicle navigates. For example, in some embodiments, the imaged objects and features may be used to locate a host vehicle relative to the map (e.g., relative to an imaged target trajectory). The imaged lane markings may be used to determine a lateral position and / or orientation with respect to a planned trajectory or target trajectory. With this position information, the autonomous vehicle may be capable of adjusting a direction of travel to correspond to a direction of a target trajectory at the determined position.The vehicle 200 may be configured to detect lane markings in a given road segment. The road segment may include any markings on a road for guiding vehicle traffic on a roadway. For example, the lane markings may be solid or dashed lines marking the edge of a lane. The lane markings may also include double lines, such as double solid lines, double dashed lines, or a combination of solid and dashed lines, indicating, for example, whether passing in an adjacent lane is permitted. The lane markings may also include highway entry and exit markings indicating, for example, a deceleration lane for an exit ramp or dotted lines indicating that a lane is turning only or that the lane ends. The markers may further indicate a working zone, a temporary lane shift, a travel path through an intersection, a median, a special lane (e.g., a bicycle lane, an HOV lane, etc.), or other various markers (e.g., pedestrian crossing, a speed ramp, a crossing, a stop line, etc.).The vehicle 200 may use cameras, such as the image capturing devices 122 and 124, included in the image capturing unit 120 to capture images of the surrounding lane markings. The vehicle 200 may analyze the images to detect point locations associated with the lane markings based on features identified within one or more of the detected images. These point locations may be uploaded to a server to represent the lane markings in sparse map 800. Depending on the position and field of view of the camera, lane markings for both sides of the vehicle may be simultaneously captured from a single image. In other embodiments, different cameras may be used to capture images on multiple sides of the vehicle. Instead of uploading actual images of the lane markings, the markings may be stored in the sparse map 800 as a spline or series of points, thereby reducing the size of the sparse map 800 and / or the data that needs to be uploaded remotely by the vehicle.FIGS. 24A-24D illustrate example point locations that may be detected by the vehicle 200 to represent particular lane markings. Similar to the landmarks described above, the vehicle 200 may use various image recognition algorithms or software to identify point locations within a captured image. For example, the vehicle 200 may recognize a series of edge points, vertices, or various other point locations associated with a particular lane marker. FIG. 24A shows a continuous lane marker 2410 that may be detected by the vehicle 200. Lane marker 2410 may represent the outer edge of a roadway, represented by a solid white line. As shown in FIG. 24A, vehicle 200 may be configured to detect a plurality of locations at edge 2411 along the lane. The location points 2411 may be collected to represent the lane marker at any intervals sufficient to create a mapped lane marker in the sparse map. For example, the lane marker may be represented by a point per meter of the detected edge, a point per every five meters of the detected edge, or at other suitable distances. In some embodiments, the distance may be determined by other factors, rather than at specified intervals, such as based on points at which the vehicle 200 has a highest confidence rating of the location of the detected points, for example. Although FIG. 24A shows the locations of the edges at an inner edge of the lane 2410, points may be collected at the outer edge of the line or along both edges. Further, although only single lines are shown in FIG. 24A, similar edges may be detected for a continuous double line. For example, the points 2411 may be detected along an edge of one or both of the solid lines.The vehicle 200 may also present lane markings differently depending on the type or shape of the lane marking. FIG. 24B shows an example dashed lane marker 2420 that may be detected by the vehicle 200. Instead of identifying the edges, as in FIG. 24A, the vehicle may detect a series of vertices 2421that represent the corners of the lane to define the full boundary of the lane. While FIG. 24B shows each corner of a particular barcode being in place, vehicle 200 may capture or upload a subset of the points shown in the figure. For example, the vehicle 200 may detect the front edge or the front corner of a given stroke mark, or may detect the two vertices closest to the inside of the lane. Further, not every barcode may be detected, for example, the vehicle 200 may detect and / or record points representing a sample of barcodes (e.g., every second, every third, every fifth, etc.) or barcodes at a predefined distance (e.g., every meter, every five meters, every 10 meters, etc.). Vertices may also be detected for similar lane markings, such as markings showing that a lane for a departure ramp is that a particular lane ends, or other various lane markings that may have detectable vertices. Vertices may also be detected for lane markings consisting of double dashed lines or a combination of continuous and dashed lines.In some embodiments, the points uploaded to the server to generate the mapped lane markings may represent other points adjacent to the detected edge points or vertices. FIG. 24C illustrates a series of points that may represent a centerline of a given lane marker. For example, the continuous lane 2410 may be represented by midline points 2441 along a midline 2440 of the lane marker. In some embodiments, the vehicle 200 may be configured to detect these midpoints using various image recognition techniques, such as convolutional neural networks (CNN), scale-invariant feature transform (SIFT), histogram of oriented gradients (HOG), or other techniques. Alternatively, vehicle 200 may detect other points, such as edge points 2411 shown in FIG. 24A, and may calculate center line points 2441, for example, by detecting points along each edge and determining a midpoint between the edge points. Similarly, dashed lane marker 2420 may be represented by midline points 2451 along a midline 2450 of the lane marker. The midline points may be located at the edge of a stroke, as shown in FIG. 24C, or at various other locations along the midline. For example, each stroke may be represented by a single point in the geometric center of the stroke. The points may also be spaced a predetermined interval along the centerline (e.g., each meter, 5 meters, 10 meters, etc.). The midline points 2451may be detected directly by the vehicle 200 or may be calculated based on other detected reference points, such as vertices 2421, as shown in FIG. 24B. A midline may also be used to represent other lane marker types, such as a double line, using techniques similar to those described above.In some embodiments, the vehicle 200 may identify points representing other features, such as an apex between two intersecting lane markings. FIG. 24D shows example points representing an intersection between two lane markings 2460 and 2465. The vehicle 200 may calculate an apex 2466 representing an intersection between the two lane markings. For example, one of the lane markings 2460 or 2465 may represent a lane transition region or another transition region in the road segment. While lane markings 2460 and 2465 are shown intersecting perpendicular to each other, various other configurations may be detected. For example, lane markings 2460 and 2465 may cross at other angles, or one or both of the lane markings may terminate at vertex 2466. Similar techniques may also be applied for intersections between dashed or other lane marker types. In addition to the vertex 2466, various other points 2467 may also be detected, which provide further information about the orientation of the lane markings 2460 and 2465.The vehicle 200 may associate real coordinates with each detected point of the lane marker. For example, location identifiers including coordinates for each point may be generated to upload to a server for mapping the lane marker. The location identifiers may further include other identifying information about the points, including whether the point represents a vertex, a border point, a midpoint, etc. The vehicle 200 may therefore be configured to determine a real position of each point based on the analysis of the images. For example, the vehicle 200 may detect other features in the image, such as the various landmarks described above, to locate the real position of the lane markings. This may include determining the location of the lane markings in the image with respect to the detected landmark or determining the position of the vehicle based on the detected landmark and then determining a distance from the vehicle (or the target trajectory of the vehicle) to the lane marking. If no landmark is available, the location of the lane markings may be determined with respect to a position of the vehicle determined based on dead reckoning navigation. The real coordinates included in the location identifiers may be represented as absolute coordinates (e.g., latitude / longitude coordinates) or may be related to other features, such as based on a longitudinal position along a target trajectory and a lateral distance from the target trajectory. The location identifiers may then be uploaded to a server to generate the mapped lane markings in the navigation model (such as sparse map 800). In some embodiments, the server may construct a spline representing the lane markings of a road segment. Alternatively, the vehicle 200 may generate the spline and upload it to the server to be recorded in the navigation model.FIG. 24E shows an example navigation model or sparse map for a corresponding road segment that includes mapped lane markings. The sparse map may include a target trajectory 2475 to be followed by a vehicle along a road segment. As described above, the target trajectory 2475 may represent an ideal path a vehicle is to take when traveling the corresponding road segment, or may be elsewhere on the road (e.g., a centerline of the road, etc.). The target trajectory 2475 may be calculated in the various methods described above, for example, based on an aggregation (e.g., a weighted combination) of two or more reconstructed trajectories of vehicles traversing the same road segment.In some embodiments, the target trajectory may be similarly generated for all vehicle types and for all road, vehicle, and / or environmental conditions. However, in other embodiments, various other factors or variables may be considered in generating the target trajectory. Another target trajectory may be generated for different vehicle types (e.g., a private vehicle, light truck, and full trailer). For example, a target trajectory with relatively narrower radii of curvature may be generated for a small private vehicle than a larger semi-trailer. In some embodiments, road, vehicle, and environmental conditions may also be considered. For example, a different target trajectory may be generated for different road conditions (e.g., wet, snow covered, icy, dry, etc.), vehicle conditions (e.g., tire condition or estimated tire condition, brake condition or estimated brake condition, remaining fuel amount, etc.), or environmental factors (e.g., time of day, visibility, weather, etc.). The target trajectory may also depend on one or more aspects or features of a particular road segment (e.g., speed limit, frequency and size of the curves, slope, etc.). In some embodiments, various user settings may also be used to determine the target trajectory, such as a fixed driving mode (e.g., desired driving aggression, economy mode, etc.).The sparse map may also include mapped lane markings 2470 and 2480, representing lane markings along the road segment. The imaged lane markings may be represented by a plurality of location identifiers 2471 and 2481. As described above, the location identifiers may include locations in real coordinates of points associated with a detected lane marker. Similar to the target trajectory in the model, the lane markings may also include height data and may be represented as a curve in three-dimensional space. For example, the curve may be a spline connecting three-dimensional polynomials of suitable order, wherein the curve may be calculated based on the location identifiers. The depicted lane markings may also include other information or metadata about the lane marking, such as an identifier of the type of lane marking (e.g., between two lanes having the same direction of travel, between two lanes having opposite directions of travel, edge of a roadway, etc.), and / or other characteristics of the lane marking (e.g., continuous, dashed, single line, double line, yellow, white, etc.). In some embodiments, the mapped lane markings within the model may be continuously updated, for example using crowdsourcing techniques. The same vehicle may upload location identifiers during multiple trips of the same road segment, or data may be selected from a plurality of vehicles (such as 1205, 1210, 1215, 1220, and 1225) that are driving the road segment at different times. Sparse map 800 may then be updated or refined based on subsequent location identifiers received from the vehicles and stored in the system. As the mapped lane markings are updated and refined, the updated road navigation model and / or sparse map may be distributed to a plurality of autonomous vehicles.Generating the imaged lane markings in the sparse map may also include detecting and / or mitigating errors based on anomalies in the images or in the actual lane markings themselves. FIG. 24F shows an example anomaly 2495 associated with detecting a lane marker 2490. The anomaly 2495 may appear in the image captured by the vehicle 200, for example, from an object that obstructs the view of the camera on the lane marking, dirt on the lens, etc. In some cases, the anomaly may be due to the lane marker itself, which may be damaged or worn or partially covered, for example, by dirt, dirt, water, snow, or other materials on the road. The anomaly 2495 may result in a faulted point 2491 being detected by the vehicle 200. Sparse map 800 may provide the correct mapped lane marker and exclude the error. In some embodiments, the vehicle 200 may detect a faulted point 2491, for example, by detecting the anomaly 2495 in the image or identifying the fault based on detected lane marker points before and after the anomaly. Based on detecting the anomaly, the vehicle may skip or adjust the point 2491 to match other detected points. In other embodiments, the error may be corrected after the point is uploaded, for example, by determining that the point is outside an expected threshold, based on other points uploaded during the same trip, or based on aggregation of previous trip data along the same road segment.The mapped lane markings in the navigation model and / or sparse map may also be used for navigation through an autonomous vehicle traversing the corresponding roadway. For example, a vehicle that navigates along a target trajectory may periodically use the mapped lane markings in the sparse map to align with the target trajectory. As mentioned above, the vehicle may navigate between landmarks based on dead reckoning, where the vehicle uses sensors to determine its own motion and estimate its position relative to the target trajectory. Errors may accumulate over time and the position determinations of the vehicle relative to the target trajectory may become more and more inaccurate. Accordingly, the vehicle may use lane markings occurring in sparse map 800 (and their known locations) to reduce dead reckoning induced errors in position determination. In this way, the identified lane markings included in sparse map 800 may serve as navigation anchors from which an accurate position of the vehicle relative to a target trajectory may be determined.FIG. 25A shows an example image 2500 of the environment of a vehicle that may be used for navigation based on the imaged lane markings. The image 2500 may be captured by, for example, the vehicle 200 via the image capturing devices 122 and 124 included in the image capturing unit 120. The image 2500 may include an image of at least one lane marker 2510, as shown in FIG. 25A. The image 2500 may also include one or more landmarks 2521, such as a road sign, used for navigation, as described above. Some elements shown in FIG. 25A, such as elements 2511, 2530, and 2520 that do not appear in the captured image 2500 but are captured and / or determined by the vehicle 200, are also shown for purposes of illustration.Using the various techniques described above with respect to FIGS. 24A-D and 24F, a vehicle may analyze the image 2500 to identify the lane marker 2510. Various points 2511 may be detected according to features of the lane marking in the image. Points 2511may correspond to, for example, an edge of the lane marking, a corner of the lane marking, a center of the lane marking, an apex between two intersecting lane markings, or various other features or locations. The points 2511 may be detected to correspond to a location of points stored in a navigation model received from a server. For example, when receiving a sparse map including points representing a center line of an imaged lane marker, the points 2511 may also be detected based on a center line of the lane marker 2510.The vehicle may also determine a longitudinal position represented by element 2520 and located along a target trajectory. The longitudinal position 2520 may be determined from the image 2500, for example, by detecting the landmark 2521 within the image 2500 and comparing a measured location to a known landmark location stored in the road model or sparse map 800. The location of the vehicle along a target trajectory may then be determined based on the distance to the landmark and the known location of the landmark. The longitudinal position 2520 may also be determined from images other than those used to determine the position of a lane marker. For example, the longitudinal position 2520 may be determined by detecting landmarks in images from other cameras within the image capturing unit 120 that are captured simultaneously or nearly simultaneously with the image 2500. In some cases, the vehicle may not be near any landmarks or other reference points for determining the longitudinal position 2520. In such cases, the vehicle may navigate based on dead reckoning and thus may use sensors to determine its own motion and estimate a longitudinal position 2520 relative to the target trajectory. The vehicle may also determine a distance 2530 that represents the actual distance between the vehicle and the lane marker 2510 observed in / in the captured image(s). The camera angle, the speed of the vehicle, the width of the vehicle, or various other factors may be taken into account in determining the distance 2530.FIG. 25B illustrates lateral localization correction of the vehicle based on the mapped lane markings in a road navigation model. As described above, the vehicle 200 may determine a distance 2530 between the vehicle 200 and a lane marker 2510 using one or more images captured by the vehicle 200. The vehicle 200 may also have access to a road navigation model, such as sparse map 800, which may include a mapped lane marker 2550 and a target trajectory 2555. The mapped lane marker 2550 may be modeled using the techniques described above, for example, using crowdsourced location identifiers captured by a plurality of vehicles. The target trajectory 2555 may also be generated using the various techniques described above. Vehicle 200 may also determine or estimate a longitudinal position 2520 along target trajectory 2555 as described above with respect to FIG. 25A. The vehicle 200 may then determine an expected distance 2540 based on a lateral distance between the target trajectory 2555 and the imaged lane marker 2550 corresponding to the longitudinal position 2520. The lateral location of the vehicle 200 may be corrected or adjusted by comparing the actual distance 2530 measured using the captured image(s) with the expected distance 2540 from the model.FIGS. 25C and 25D provide illustrations associated with another example of locating a host vehicle during navigation based on mapped landmarks / objects / features in a sparse map. FIG. 25C conceptually illustrates a series of images captured by a vehicle that navigates along a road segment 2560. In this example, road segment 2560 includes a straight portion of a two-lane split highway bounded by road edges 2561 and 2562 and center lane marker 2563. As shown, the host vehicle navigates along a lane 2564 associated with an imaged target trajectory 2565. Therefore, in an ideal situation (and without impact factors such as the presence of target vehicles or objects on the roadway, etc.), the host vehicle should accurately follow the mapped target trajectory 2565 as it navigates along the lane 2564 of the road segment 2560. In reality, the host vehicle may experience a drift as it navigates along the mapped target trajectory 2565. For effective and safe navigation, this drift should be kept within acceptable limits (e.g., + / - 10 cm lateral displacement from the target trajectory 2565, or any other suitable threshold). To periodically account for the drift and make all required heading corrections to ensure that the host vehicle follows the target trajectory 2565, the disclosed navigation systems may be capable of locating the host vehicle along the target trajectory 2565 (e.g., determining a lateral and longitudinal position of the host vehicle relative to the target trajectory 2565) using one or more mapped features / objects included in the sparse map.As a simple example, FIG. 25C shows a speed limit sign 2566 as it may appear in five different images captured in succession as the host vehicle travels along the road segment 2560. For example, at a first time, t0, the sign 2566 may appear in a captured image near the horizon. As the host vehicle approaches the sign 2566, in the subsequent captured images at times t 1, t 2, t 3, and t 4, the sign 2566 appears at different 2D-X-Y pixel positions of the captured images. For example, in the space of the captured images, the sign 2566 moves down and to the right along the curve 2567 (e.g., a curve that extends through the center of the sign in each of the five captured images). The sign 2566 also appears to increase in size as the host vehicle approaches it (i.e., it occupies a large number of pixels in the subsequently captured images).These changes in the image space representations of an object, such as the sign 2566, may be exploited to determine the local position of the host vehicle along a target trajectory. For example, as described in the present disclosure, any detectable object or characteristic, such as a semantic characteristic such as the sign 2566 or a detectable non-semantic characteristic, may be identified by one or more crop vehicles that have previously traversed a road segment (e.g., the road segment 2560). A mapping server may collect the collected driving information from a plurality of vehicles, aggregate and correlate this information, and generate a sparse map including, for example, a target trajectory 2565 for the lane 2564 of the road segment 2560. The sparse map may also store the location of the sign 2566 (along with type information, etc.). During navigation (e.g., prior to entering road segment 2560), a host vehicle may be provided with a map tile that includes a sparse map for road segment 2560. To navigate the lane 2564 of the road segment 2560, the host vehicle may follow the mapped target trajectory 2565.The mapped representation of the sign 2566 may be used by the host vehicle to locate itself relative to the target trajectory. For example, a camera on the host vehicle captures an image 2570 of the environment of the host vehicle, and this captured image 2570 may include an image representation of the sign 2566 having a particular size and X-Y image position, as shown in FIG. 25D. This size and X-Y image position may be used to determine the position of the host vehicle relative to the target trajectory 2565. For example, a navigation processor of the host vehicle may determine, based on the sparse map including a representation of the sign 2566, that in response to the host vehicle traveling along the target trajectory 2565, a representation of the sign 2566 should appear in the captured images such that a center of the sign 2566 (in image space) moves along the line 2567. If a captured image, such as image 2570, shows that the center point (or other reference point) is displaced from the line 2567 (e.g., the expected image space trajectory), the host vehicle navigation system may determine that it was not on the target trajectory 2565 at the time of the captured image. However, from the image, the navigation processor may determine an appropriate navigation correction to return the host vehicle to the target trajectory 2565. For example, if the analysis shows an image position of the sign 2566 that is left-shifted in the image by a distance 2572 from the expected image space position on the line 2567, the navigation processor may cause a turn of the host vehicle to move (e.g., by changing the steering angle of the wheels) to the left of the host vehicle by a distance 2573. In this way, each captured image may be used as part of a feedback loop such that a difference between an observed image position of the sign 2566 and the expected image trajectory 2567 may be minimized to ensure that the host vehicle continues along the target trajectory 2565 with little or no deviation. Of course, the localization technique described can be used more frequently the more imaged objects are available, whereby deviations from the target trajectory 2565, caused by drift, can be reduced or eliminated.The process described above may be useful for detecting a lateral orientation or displacement of the host vehicle relative to a target trajectory. Locating the host vehicle relative to the target trajectory 2565 may also include determining the longitudinal position of the target vehicle along the target trajectory. For example, the captured image 2570 includes a representation of the sign 2566 having a particular image size (e.g., 2D X Y pixel area). This size may be compared to an expected image size of the imaged sign 2566 as it travels along line 2567 through the image space (e.g., as the size of the sign progressively increases, as shown in FIG. 25C ). Based on the image size of the sign 2566 in image 2570 and the expected size profile in image space relative to the mapped target trajectory 2565, the host vehicle may determine its longitudinal position (at the time of acquiring image 2570) relative to the target trajectory 2565. This longitudinal position, coupled with any lateral displacement relative to the target trajectory 2565, as described above, allows for full localization of the host vehicle relative to the target trajectory 2565 as the host vehicle navigates along the road 2560.FIGS. 25C and 25D depict only one example of the disclosed localization technique using a single imaged object and a single target trajectory. In other examples, there may be many more target trajectories (e.g., a target trajectory for each usable lane of a multi-lane highway, urban road, complex intersection, etc.) and there may be many more imaged objects for localization. For example, a sparse map representative of an urban environment may include many objects per meter available for localization.FIG. 26A is a flow diagram illustrating an example process 2600A for mapping a lane marker for use in autonomous vehicle navigation, in accordance with the disclosed embodiments. At step 2610, process 2600A may include receiving two or more location identifiers associated with a detected lane marker. For example, step 2610 may be performed by the server 1230 or one or more processors associated with the server. The identifications of locations may include locations in real coordinates of points associated with the detected lane, as described above with respect to FIG. 24E. In some embodiments, the location identifiers may also include other data, such as additional information about the road segment or lane marker. Additional data may also be received during step 2610, such as accelerometer data, speed data, landmark data, road geometry or profile data, vehicle positioning data, ego motion data, or various other forms of data described above. The location identifiers may be generated by a vehicle, such as vehicles 1205, 1210, 1215, 1220, and 1225 based on images captured by the vehicle. For example, the identifiers may be determined based on the acquisition of at least one image representing an environment of the host vehicle from a camera associated with a host vehicle, the analysis of the at least one image to detect the lane marker in the environment of the host vehicle, and the analysis of the at least one image to determine a position of the detected lane marker relative to a location associated with the host vehicle. As described above, the lane marking may include a plurality of different types of markings, and the location identifiers may correspond to a plurality of points relative to the lane marking. For example, if the detected lane marker is part of a dashed line marking a lane boundary, the points may correspond to detected corners of the lane marker. When the detected lane marking is part of a solid line marking a lane boundary, the points may correspond to a detected edge of the lane marking at various distances as described above. In some embodiments, the points may correspond to the centerline of the detected lane marker as shown in FIG. 24C, or may correspond to an apex between two intersecting lane markers and at least one or two other points associated with the intersecting lane markers as shown in FIG. 24D.At step 2612, process 2600A may include associating the detected lane marker with a corresponding road segment. For example, the server 1230 may analyze the real coordinates or other information received during step 2610, and compare the coordinates or other information to location information stored in a road navigation model of the autonomous vehicle. The server 1230 may determine a road segment in the model corresponding to the real road segment in which the lane marker was detected.At step 2614, process 2600A may include updating a road navigation model of the autonomous vehicle relative to the corresponding road segment based on the two or more location identifiers associated with the detected lane marker. For example, the autonomous road navigation model may be sparse map 800 and server 1230 may update the sparse map to include or fit a mapped lane marker into the model. Server 1230 may update the model based on the various methods or processes described above with reference to FIG. 24E. In some embodiments, updating the autonomous vehicle road navigation model may include storing one or more position indicator(s) in real coordinates of the detected lane marker. The autonomous vehicle road navigation model may also include at least one target trajectory to be followed by a vehicle along the corresponding road segment, as shown in FIG. 24E.At step 2616, process 2600A may include distributing the updated autonomous vehicle road navigation model to a plurality of autonomous vehicles. For example, the server 1230 may distribute the updated autonomous vehicle road navigation model to the vehicles 1205, 1210, 1215, 1220, and 1225, which may use the model for navigation. The autonomous vehicle road navigation model may be distributed over one or more networks (e.g., over a cellular network and / or the Internet, etc.) via wireless communication paths 1235, as shown in FIG. 12.In some embodiments, the lane markings may be mapped using data received from a plurality of vehicles, such as by a crowdsourcing technique as described above with respect to FIG. 24E. For example, process 2600A may include receiving a first communication from a first host vehicle including location identifiers associated with a detected lane marker and receiving a second communication from a second host vehicle including additional location identifiers associated with the detected lane marker. For example, the second communication may be received from a trailing vehicle traveling on the same road segment or from the same vehicle on a trailing trip along the same road segment. The process 2600A may further include refining a determination of at least one position associated with the detected lane marker based on the location identifiers received in the first communication and based on the additional location identifiers received in the second communication. This may include using an average of the plurality of location identifiers and / or filtering out ghost identifiers that may not reflect the real position of the lane marker.FIG. 26B is a flowchart illustrating an example process 2600B for autonomously navigating a host vehicle along a road segment using mapped lane markings. The process 2600B may be performed by the processing unit 110 of the autonomous vehicle 200, for example. At step 2620, process 2600B may include receiving a road navigation model of the autonomous vehicle from a server-based system. In some embodiments, the autonomous vehicle road navigation model may include a target trajectory for the host vehicle along the road segment and location identifiers associated with one or more lane markings associated with the road segment. For example, the vehicle 200 may receive a sparse map 800 or other road navigation model developed using process 2600A. In some embodiments, the target trajectory may be represented as a three-dimensional spline, for example, as shown in FIG. 9B. As described above with respect to FIGS. 24A-F, the location identifiers may include locations in real coordinates of points associated with the lane marker (e.g., vertices of a dashed lane marker, edge points of a continuous lane marker, a vertex between two intersecting lane markers and other points associated with the intersecting lane markers, a centerline associated with the lane marker, etc.).At step 2621, process 2600B may include receiving at least one image representing an environment of the vehicle. The image may be received by an image capture device of the vehicle, such as by the image capture devices 122 and 124 included in the image capture unit 120. The image may include an image of one or more lane markings, similar to image 2500 described above.At step 2622, the process 2600B may include determining a longitudinal position of the host vehicle along the target trajectory. As described above with respect to FIG. 25A, this may be based on other information in the captured image (e.g., landmarks, etc.) or on dead reckoning of the vehicle between detected landmarks.At step 2623, process 2600B may include determining an expected lateral distance to the lane marker based on the determined longitudinal position of the host vehicle along the target trajectory and based on the two or more location identifiers associated with the at least one lane marker. For example, the vehicle 200 may use the sparse map 800 to determine an expected lateral distance to the lane marker. As shown in FIG. 25B, the longitudinal position 2520 along a target trajectory 2555 may be determined in step 2622. Using replacement map 800, vehicle 200 may determine an expected distance 2540 to imaged lane marker 2550 corresponding to longitudinal position 2520.At step 2624, the process 2600B may include analyzing the at least one image to identify the at least one lane marker. For example, the vehicle 200 may use various image recognition techniques or algorithms to identify the lane marker within the image, as described above. For example, lane marker 2510 may be detected by image analysis of image 2500, as shown in FIG. 25A.At step 2625, process 2600B may include determining an actual lateral distance to the at least one lane marker based on the analysis of the at least one image. For example, the vehicle may determine a distance 2530, as shown in FIG. 25A, which represents the actual distance between the vehicle and the lane 2510. The camera angle, the speed of the vehicle, the width of the vehicle, the position of the camera relative to the vehicle, or various other factors may be taken into account in determining the distance 2530.At step 2626, process 2600B may include determining an autonomous steering action for the host vehicle based on a difference between the expected lateral distance to the at least one lane marker and the determined actual lateral distance to the at least one lane marker. For example, as described above with respect to FIG. 25B, vehicle 200 may compare actual distance 2530 to expected distance 2540. The difference between the actual and expected distances may indicate an error (and its magnitude) between the actual position of the vehicle and the target trajectory to be tracked by the vehicle. Accordingly, the vehicle may determine an autonomous steering action or other autonomous action based on the difference. For example, if the actual distance 2530 is less than the expected distance 2540 as shown in FIG. 25B, the vehicle may determine an autonomous steering motion to steer the vehicle to the left, away from the lane marker 2510. Thus, the position of the vehicle with respect to the target trajectory may be corrected. For example, process 2600B may be used to improve navigation of the vehicle between landmarks.Processes 2600A and 2600B provide only examples of techniques that may be used to navigate a host vehicle using the disclosed sparse maps. In other examples, consistent processes described with those described with reference to FIGS. 25C and 25D may also be employed.In some embodiments, the disclosed systems, methods, and non-transitory computer readable media may use one or more AV cards. An autonomous vehicle (AV) (or AV) map may include information for supporting and / or implementing one or more functions of a vehicle in a manner that the vehicle is safely operated and / or navigates precisely. The AV functions supported and / or implemented by an autonomous vehicle (or a semi-autonomous vehicle) may include one or more autonomously controlled functions (e.g., functions determined, selected, and / or implemented based on instructions executed by at least one processor), such as steering, accelerating, and / or braking the vehicle. The AV functions may be part of a driving style, such as RSS developed and implemented by Mobileye, Jerusalem, Israel. The information for safe operation of the vehicle may include, but is not limited to, information related to one or more regulations applicable to a location of the vehicle or a jurisdiction (e.g., a jurisdiction with right or left traffic), information related to an environment in which the vehicle is located (e.g., information related to a drivable path, a stop sign, a traffic light, a speed limit, lane markings, landmarks, free space, virtual or physical stop lines, information related to traffic lights relevance, etc.), and / or information related to adjusting and / or tuning navigation of the vehicle to include one or more objects (e.g., other vehicles, pedestrians, objects, This is because obstacles, occlusions, hazards, construction sites, traffic cones, etc.) in the surroundings of the vehicle are to be taken into account. The vehicle may detect objects in the environment of the vehicle using one or more sensors (e.g., cameras, radar, lidar), as discussed herein. The AV map may serve as a redundant information source for the captured information and, in some cases, may also supplement the captured information (e.g., provide the location of a virtual stop line when no stop line is marked on the road). In some embodiments, safely guiding the vehicle may further include guiding the vehicle to maintain a level of comfort for one or more passengers in the vehicle. Comfort may include one or more predetermined criteria (e.g., in terms of speed, acceleration, and / or cornering) selected to operate the vehicle to reach or exceed an intended or selected level of passenger comfort. The level of comfort may be formalized and expressed in suitable mathematical formulas. For example, mathematical formulas can be used that limit the amount of acceleration or jerk applied to a passenger in various directions. The information for proper operation of the vehicle may include, but is not limited to, information relating to a planned or intended path (e.g., a trajectory as described herein) or route (e.g., from a particular location, such as a starting location, to a destination location). In some embodiments, the information included in an AV card may further include information that supports and / or implements one or more AV functions in an efficient manner. The information for efficient operation of the vehicle may include speed, acceleration, lane change, and / or positioning information in the lane and / or driving, and / or selection of a path or route based on traffic conditions (e.g., driving a route that is longer than a shorter route with traffic conditions) and / or other factors or attributes associated with potential routes (e.g., weather conditions, road condition, or other characteristics of routes, such as travel on a highway without traffic lights instead of a road with traffic lights) In some embodiments, the AV map may further include at least some information from a high resolution (HD) map. In some embodiments, the AV card may be a sparse card as described above. In some embodiments, the AV map may be created by crowdsourcing as discussed herein. In this disclosure, the terms "AV map" and "sparse map" are used interchangeably.Map-Based Detection of ConditionA host vehicle (autonomous or semi-autonomous) may include cameras that capture images from the environment of the host vehicle, and the host vehicle may include a navigation system that analyzes the captured images to make navigation decisions. Current navigation systems typically capture high resolution images and reduce their resolution so that the systems can process the images more quickly and with less computational resources. However, this will lose information and accuracy that was present in the original high resolution images. The use of low resolution images may present a challenge when the images include representations of remote objects. Because the remote objects are far from the cameras and the images are of low resolution, it is often difficult for the navigation systems to identify the objects themselves as well as the regions of the images that include the representations of the remote objects. The inability to identify regions in the image that include representations of remote objects may result in the remote objects being misidentified. Alternatively, if the navigation systems can locate these regions in the image, they can use an augmented version of the images to analyze these regions in the image for potential objects. However, even if these regions have been found, it is a challenge for the navigation systems to align the relative position of the zoomed images to the position of the host vehicle. Accordingly, although the navigation system may identify a remote object, it may not correctly associate the location of the remote object with the location of the vehicle. The disclosed embodiments include systems and methods that improve these existing techniques and mitigate or eliminate the foregoing disadvantages.FIG. 27 is a flow diagram illustrating an example process 2700 for navigation of a host vehicle consistent with the disclosed embodiments. Process 2700 may be performed by at least one processing device, such as processing unit 110 included in system 100, or various other devices described herein. It is understood that the term "processor" or "processor unit" is used in the above and the present specification as an abbreviation for "at least one processor" or "at least one processor unit". In other words, a processor or processing unit may include one or more structures (e.g., circuitry) that perform logical operations regardless of whether these structures are collapsed, connected, or dispersed. In some embodiments, a non-transitory computer readable medium may include instructions that, when executed by a processor, cause the processor to perform process 2700. Further, the process 2700 is not necessarily limited to the steps shown in FIG. 27, at least some of the steps shown in FIG. 27 may be optional, and all steps or processes of the various embodiments described in the present disclosure may also be included in the process 2700.At step 2702, the processing unit 110 may receive at least one image captured by a camera of the host vehicle from an environment of the host vehicle. For example, a camera included in the image capture unit 120 (such as the image capture device 122, 124, or 126, each having fields of view 202, 204, and 206) of the system 100 may capture at least one image of an area around the host vehicle 200 and transmit it to the processing unit 110 via a connection (e.g., digital, wired, USB, wireless, Bluetooth, etc.) via a data interface 128. In some embodiments, the camera may be a front camera of the host vehicle, such as camera 122, that captures an area in front of the host vehicle 200. In some other embodiments, the processing unit 110 may receive a plurality of captured images from one or more cameras onboard the host vehicle. For example, image capture devices 122, 124, and 126 having fields of view 202, 204, and 206, respectively, included in the image capture unit 120 may transmit all images representative of the environment of the host vehicle. Additionally, in some embodiments, the at least one captured image may include a representation of at least one object in the environment of the host vehicle. FIG. 28 illustrates an example image 2800 captured by the camera 122 of the host vehicle 200 and illustrating the vicinity thereof. In image 2800, a road segment 2810 is shown on which the host vehicle 200 is traveling. In addition, the image includes representations of road markings 2815 associated with the road segment 2810, as well as various objects in the environment of the host vehicle, such as another vehicle 2830 and a tree 2850.At step 2704, the processing unit 110 may analyze the at least one image to identify a region of interest in the environment of the host vehicle. As used herein, a region of interest (ROI) in the environment of a host vehicle refers to an area or zone of particular importance to the systems of the host vehicle. A region of interest may be determined based on factors requiring attention of the host vehicle, such as obstacles or objects, road boundaries, or traffic signs. In the context of the perception system of a host vehicle, ROls may help concentrate the energy and sensors of the processor to relevant areas, thus improving safety, efficiency, and situational awareness. In some embodiments, the region of interest in the environment of the host vehicle may be identified based on a trajectory or route of the host vehicle. In still other embodiments, a region of interest may be determined based on one or more outputs provided by a trained system (e.g., a trained network such as a neural network). For example, the trained system may be configured to learn to identify one or more regions of interest in one or more images based on one or more of the above factors and then output one or more portions of the one or more images associated with the identified region(s) of interest. A route refers to a predefined or planned path that the host vehicle wishes to follow to reach a destination, taking into account the general position and direction of the vehicle. For example, a route may include a series of instructions provided by a navigation system or application of the host vehicle followed by the host vehicle to travel from a first location (e.g., an origin location) to a second location (e.g., a destination location). The instructions may identify particular roads, locations, landmarks, etc., that together include the route and the relationships (e.g., directions) between the roads, locations, landmarks, etc., along the route. A trajectory not only follows the planned route, but also takes into account the real-time dynamics of the vehicle, such as its speed, its exact position, and any adjustments required along the path to bypass obstacles or changes in the environment. Information about the host vehicle's route or trajectory, i.e., details regarding the intended direction of the host vehicle, may help predict where the host vehicle will be in the future based on captured images of the environment. This predictive capability may enable the identification of areas in the environment that are potentially potentially hazardous or require attention of the host vehicle, i.e., regions of interest. That is, a trajectory may represent a positioning of the host vehicle along one or more roads along a route (e.g., a three-dimensional spline that identifies a drivable path). In some embodiments, reference to a trajectory may facilitate identifying a more accurate region of interest. Accordingly, in some embodiments, the route-based identification of a region of interest may be less accurate than the trajectory-based identification of a rejection of interest, and the route-based identification may therefore use additional information to identify the region of interest. Referring to FIG. 28, the provision of the trajectory or route 2840 of the host vehicle 200 (shown as a black dashed line) may assist the processing unit 110 to identify a region of interest within the image 2800. In the scenario depicted in FIG. 28, the region of interest is in an area in front of the host vehicle 200.At step 2706, the processing unit 110 may select a portion of the at least one image based on the region of interest. For example, referring to FIG. 28, the processing unit 110 may select the portion 2820 of the image 2800 based on the region of interest identified as being in a region in front of the host vehicle 200. The portion of the at least one image may include, for example, a certain percentage of the image (e.g., 50%, 25%, 10%, etc.). Additionally, in some embodiments, the processing unit 110 may cut the selected portion from the original image. The cut portion may maintain the resolution of at least one image (e.g., maintain a high resolution).At step 2708, the process may receive map information associated with the environment of the host vehicle. The map information may include one or more identifiers of a route of the host vehicle. As used herein, map information associated with the environment of a host vehicle refers to geographic and spatial data relevant to the immediate environment of the host vehicle. Map information may include details such as road shapes, intersections, traffic signs, lane markings, speed limits, a drawn or virtual stop line, and nearby obstacles or landmarks. In accordance with the disclosed embodiments, map information may include one or more identifications of the route of the host vehicle.In some embodiments, the one or more identifications of the host vehicle's route may be associated with a road segment. For example, one or more identifications may include GPS coordinates that track the location of the host vehicle along the road segment, markings of intersections or decision points on the road segment, or entry and exit points that divide the road segment into sections. In some embodiments, the one or more identifiers of the host vehicle's route may be associated with a lane of a road segment. For example, one or more identifications may include position in the lane, lane type, lane width, lane curvature, or lane identification number to distinguish lanes on multi-lane roads. In addition, data on lane markings, such as solid or dashed lines, may be included to indicate lane boundaries, merging areas, or exit lanes. Additionally or alternatively, in some embodiments, the one or more identifications of the route of the host vehicle relative to a center of the lane. Such data may assist in lane keeping assistance by guiding the host vehicle on the optimal path for safety and stability (e.g., by tracking the lateral distance from the lane center). By combining these various kinds of identifications, comprehensive map information can be obtained. For example, in some embodiments, the map information may include at least one of a lane, an edge, a lane marker, and a travelable path.In some embodiments, the region of interest in the environment of the host vehicle may be identified based on the map information. In other words, the processing unit 110 may use the received map information (e.g., the location of the lane at a meter distance, the location of one or more edges or markings of roads, etc.) to determine potential regions of interest in at least one captured image.At step 2710, the processing unit 110 may provide the portion of the at least one image and the map information to a trained system. The trained system may be configured to generate one or more detections using a language model architecture based on the analysis of the portion of the at least one image and the map information. It should be appreciated that in alternative embodiments, the processing unit 110 may provide the at least one image (the entire captured image) and the map information to a trained system. In other words, the steps of image analysis to identify an ROI and selecting / cropping a portion of the image based on the identified ROI as described in steps 2704 and 2706 are optional. The trained system may be configured to generate one or more predictions using a language model architecture based on analysis of the at least one image and the map information.Herein, a trained system refers to a computing system or algorithm that has undergone a training process using a dataset to learn patterns, features, or relationships within the data. Training may include adjusting the system parameters through techniques such as supervised, unsupervised, or reinforcement learning. Training may include adjusting the system parameters through techniques such as supervised, unsupervised, or reinforcement learning. In some embodiments, the trained system may include one or more machine learning models, one or more neural networks, and / or one or more deep learning architectures. In some embodiments, the trained system may include one or more machine learning models, one or more neural networks, and / or one or more deep learning architectures. In some examples, an artificial neural network (e.g., a deep neural network, a convolutional neural network, etc.) may be configured using the training data. The training data may include a series of images and a series of indicators or labels corresponding to areas of images that include remote objects.Trained systems can be used in various applications, such as image recognition, natural language processing, and autonomous systems. The term hardware implementation is the range of possible hardware components for trained systems, ranging from general purpose processors to fully customized hardware chips. Examples of hardware components used to implement a trained neural network may be CPUs (central processing units), GPUs (graphics processors), ASICs (application specific integrated circuits), FPGAs (field programmable gate arrays), CMOS microcontrollers, neurochips or neurocomputers. In some embodiments, the trained system may be implemented directly on system 100 (e.g., in processing unit 110). Alternatively, in some other embodiments, the at least one trained neural network may be external to the system 110 and the processing unit 110 may communicate with the trained neural system (i.e., input of one or more different data inputs to the trained system and / or reading or extracting one or more outputs from the trained system) using any known means for sending and / or receiving data. For example, the processing unit 110 may use the wireless transceiver 172 to interact and communicate with an external, trained system.In accordance with the disclosed embodiments, the trained system may be configured to generate one or more predictions using a language model architecture. In this document, the term "language model architecture" refers to the structured design and framework of a model specifically designed for understanding, generating, and editing human speech. Such an architecture defines how the model processes text data, learns speech patterns, and generates meaningful outputs. Examples of types of speech model architectures include recurrent neural networks (RNNs), long-term short-term memory (LSTM), transformers, BERT (bidirectional encoder representations from transformers), or GPT (generative pre-trained transformers). Although these models have been developed in the past and have evolved into de facto standards for natural language processing (NLP) tasks such as machine translation, text summary, mood analysis, or conversation AI, their architectures are suitable for computer vision tasks. The flexibility of these architectures relies largely on the basic common principles of understanding patterns and relationships in data, whether text or image data. For example, the transformer architecture has been successfully applied to computer vision by models such as vision transformer (ViT). In this adaptation, images are treated as sequences of patches, similar to words in speech models are treated as sequences of tokens. This allows the model to capture long range dependencies and contextual information within the i...

Claims

A system for navigating a host vehicle, the system comprising: at least one processor comprising a circuit and a memory, the memory including instructions that, when executed by the circuit, cause the at least one processor to: receive at least one image captured by a camera of the host vehicle from an environment of the host vehicle; analyze the at least one image to identify an area of interest in the environment of the host vehicle; select a portion of the at least one image based on the region of interest; receive map information associated with the environment of the host vehicle, wherein the map information includes one or more identifications of a route of the host vehicle; providing the portion of the at least one image and the map information to a trained system, the trained system configured to generate one or more detections using a language model architecture based on the analysis of the portion of the at least one image and the map information; receiving an output provided by the trained system, the output including an identification of an object in the environment of the host vehicle and location information for the object relative to the map information; and causing the host vehicle to initiate at least one navigation action based on the identification of the object and the information about the location of the object.The system of claim 1, wherein the camera is a front view camera of the host vehicle.The system of claim 1, wherein the at least one image includes a representation of the object.The system of claim 1, wherein the region of interest in the environment of the host vehicle is identified based on a trajectory of the host vehicle.The system of claim 1, wherein the region of interest in the environment of the host vehicle is identified based on the route of the host vehicle.The system of claim 1, wherein the region of interest in the environment of the host vehicle is identified based on the map information.The system of claim 1, wherein the one or more identifications of the host vehicle's route are associated with a road segment.The system of claim 1, wherein the one or more identifications of the route of the host vehicle are associated with a lane of a road segment.The system of claim 1, wherein the one or more identifications of the host vehicle's route relate to a lane center.The system of claim 1, wherein the map information includes at least one of a lane center, a road edge, a lane marker, and a travelable path.The system of claim 1, wherein the location information for the object includes coordinate information.The system of claim 1, wherein the location information for the object relates to the route of the host vehicle.The system of claim 1, wherein the trained system is further configured to identify the object in the environment of the host vehicle.The system of claim 1, wherein the trained system is further configured to generate information indicative of the location information for the object relative to the map information.The system of claim 1, wherein the trained system is further configured to generate a sequence of tokens representing the object based on the portion of the at least one image.The system of claim 15, wherein the trained system is further configured to determine an identification of the object in the environment of the host vehicle based on the sequence of tokens.The system of claim 1, wherein the trained system includes one or more machine learning models.The system of claim 1, wherein the trained system includes one or more neural networks.The system of claim 1, wherein the identification of the object includes a category of the object.The system of claim 19, wherein the category of the object includes a target vehicle or a pedestrian.The system of claim 1, wherein the location information for the object includes the coordinates of the object.The system of claim 1, wherein the object is a target vehicle.The system of claim 1, wherein the object is a pedestrian.The system of claim 1, wherein the at least one navigation action includes deceleration, acceleration, or steering of the host vehicle.The system of claim 1, wherein the memory further includes instructions that, when executed by the circuitry, cause the at least one processor to provide images selected from the captured images to the trained system.The system of claim 24, wherein the selected images have a lower resolution than the resolution of the selected portion of the at least one image.The system of claim 1, wherein the identification comprises a unique identifier of the detected object.The system of claim 27, wherein the unique identifier of the object is used to track the object between sequences of detections.The system of claim 1, wherein the language model is configured to generate a language-based description of the scene.The system of claim 29, wherein the description comprises at least one detected object, wherein the description comprises at least one of a location, a category, and an identification of the object.A method of navigating a host vehicle, the method comprising: receiving at least one image captured by a camera of the host vehicle from an environment of the host vehicle; analyzing the at least one image to identify an area of interest in the environment of the host vehicle; selecting a portion of the at least one image based on the region of interest; receiving map information associated with the environment of the host vehicle, wherein the map information includes one or more identifications of a route of the host vehicle; providing the portion of the at least one image and the map information to a trained system, wherein the trained system is configured to generate one or more detections using a language model architecture based on the analysis of the portion of the at least one image and the map information; receiving an output provided by the trained system, the output including an identification of an object in the environment of the host vehicle and location information for the object relative to the map information; and causing the host vehicle to initiate at least one navigation action based on the identification of the object and the information about the location of the object.The method of claim 31, wherein the region of interest in the environment of the host vehicle is identified based on a trajectory of the host vehicle.The method of claim 31, wherein the region of interest in the environment of the host vehicle is identified based on the route of the host vehicle.The method of claim 31, wherein the region of interest in the environment of the host vehicle is identified based on the map information.The method of claim 31, wherein the method further comprises providing images selected from the captured images to the trained system.The method of claim 35, wherein the selected images have a lower resolution than the resolution of the selected portion of the at least one image.The method of claim 31, wherein the language model is configured to generate a language-based description of the scene.The method of claim 37, wherein the description comprises at least one detected object, wherein the description comprises at least one of a location, a category, and an identification of the object.A non-transitory computer readable medium storing program instructions executable by at least one processor to perform a method, the method comprising: receiving at least one image captured by a camera of the host vehicle from an environment of the host vehicle; analyzing the at least one image to identify an area of interest in the environment of the host vehicle; selecting a portion of the at least one image based on the region of interest; receiving map information associated with the environment of the host vehicle, wherein the map information includes one or more identifications of a route of the host vehicle; providing the portion of the at least one image and the map information to a trained system, the trained system configured to generate one or more detections using a language model architecture based on the analysis of the portion of the at least one image and the map information; receiving an output provided by the trained system, the output including an identification of an object in the environment of the host vehicle and location information for the object relative to the map information; and causing the host vehicle to initiate at least one navigation action based on the identification of the object and the information about the location of the object.The non-transitory computer readable medium of claim 39, wherein the region of interest in the environment of the host vehicle is identified based on a trajectory of the host vehicle.The non-transitory computer readable medium of claim 39, wherein the region of interest in the environment of the host vehicle is identified based on the route of the host vehicle.The non-transitory computer readable medium of claim 39, wherein the region of interest in the environment of the host vehicle is identified based on the map information.The non-transitory computer readable medium of claim 39, wherein the language model is configured to generate a language-based description of the scene.A non-transitory computer readable medium storing program instructions executable by at least one processor to perform a method, the method comprising: receiving at least one image captured by a camera of the host vehicle from an environment of the host vehicle, analyzing the at least one image to identify an area of interest in the environment of the host vehicle; selecting a portion of the at least one image based on the region of interest; receiving map information associated with the environment of the host vehicle, wherein the map information includes one or more identifications of a route of the host vehicle; providing the portion of the at least one image and the map information to a trained system, the trained system configured to generate one or more predictions using a language model architecture based on the analysis of the portion of the at least one image and the map information; receiving an output provided by the trained system, the output including an identification of an object in the environment of the host vehicle and location information for the object relative to the map information; and causing the host vehicle to initiate at least one navigation action based on the identification of the object and the information about the location of the object.A non-transitory computer readable medium storing program instructions executable by at least one processor to perform a method, the method comprising: receiving at least one image captured by a camera of the host vehicle from an environment of the host vehicle; receiving map information associated with the environment of the host vehicle, the map information including one or more identifications of a route of the host vehicle; providing the portion of the at least one image and the map information to a trained system, wherein the trained system is configured to generate one or more predictions using a language model architecture based on the analysis of the portion of the at least one image and the map information; receiving an output provided by the trained system, the output including an identification of an object in the environment of the host vehicle and location information for the object relative to the map information; and causing the host vehicle to initiate at least one navigation action based on the identification of the object and the information about the location of the object.

Images