High-precision map element vectorization method and device, electronic equipment and storage medium

Abstract

The present disclosure provides a map element vectorization method and device, electronic equipment, storage medium and product, relates to the positioning technology field of an autonomous vehicle, and in particular to the positioning technology field of vectorizing element features based on a high-precision map. The specific implementation scheme is as follows: an image of a road is acquired, and point cloud data of the road is acquired; a longitudinal structure object in the image is detected, a detection position and an attribute of the longitudinal structure object are determined; and based on the detection position and the attribute, a feature of the longitudinal structure object in the point cloud data is determined. Through the present disclosure, the recall rate of the feature of the longitudinal structure object in the high-precision map can be improved.

Description

Technical Field

[0001] This disclosure relates to the field of autonomous vehicle positioning technology, and more particularly to the field of high-precision map vectorized feature positioning technology, specifically to a map feature vectorization method, device, electronic device, storage medium, and product. Background Technology

[0002] High-precision maps, also known as high-resolution maps, are used by autonomous vehicles. They possess accurate vehicle location information and rich road element data, helping cars anticipate complex road conditions such as slope, curvature, and heading, thus better avoiding potential risks. The positioning layer in high-precision maps provides basic map positioning services through lane lines, signs, and poles, especially in autonomous vehicle positioning products (Apollo Navigation Pilot, ANP).

[0003] In related technologies, ANP's positioning layer is based on lane lines for positioning, which leads to significant deviations in the positioning of autonomous vehicles on the road. Summary of the Invention

[0004] This disclosure provides a method, apparatus, electronic device, storage medium, and product for vectorizing map features.

[0005] According to a first aspect of this disclosure, a method for determining map features is provided, the method comprising:

[0006] Acquire an image of the road and obtain point cloud data of the road; detect longitudinal structural objects in the image and determine the detection position and attributes of the longitudinal structural objects; based on the detection position and attributes, determine the features of the longitudinal structural objects in the point cloud data.

[0007] According to a second aspect of this disclosure, a map feature determination apparatus is provided, the apparatus comprising:

[0008] An acquisition module is used to acquire an image of a road and acquire point cloud data of the road; a determination module is used to detect vertical structural objects in the image and determine the detection position and attributes of the vertical structural objects; the determination module is also used to determine the features of the vertical structural objects in the point cloud data based on the detection position and attributes.

[0009] According to a third aspect of this disclosure, an electronic device is provided, comprising:

[0010] At least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of the first aspect.

[0011] According to a fourth aspect of this disclosure, a non-transitory computer-readable storage medium is provided storing computer instructions, wherein the computer instructions are configured to cause the computer to perform the method according to the first aspect.

[0012] According to a fifth aspect of this disclosure, a computer product is provided, including a computer program that, when executed by a processor, implements the method according to the first aspect.

[0013] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0014] The accompanying drawings are provided to better understand this solution and do not constitute a limitation of this disclosure. Wherein:

[0015] Figure 1 This is a schematic diagram illustrating an application scenario of an embodiment of this disclosure;

[0016] Figure 2 A schematic flowchart of a map feature determination method provided in an embodiment of this disclosure is shown;

[0017] Figure 3 A flowchart illustrating a method for determining the detection position and attributes of a longitudinally structured object according to an embodiment of this disclosure is shown.

[0018] Figure 4 A flowchart illustrating a positioning layer generation method provided in an embodiment of this disclosure is shown.

[0019] Figure 5 A flowchart illustrating a method for determining the features of a longitudinally structured object according to an embodiment of this disclosure is shown.

[0020] Figure 6 A schematic diagram of the sign vectorization provided in this disclosure embodiment is shown;

[0021] Figure 7 A schematic diagram of rod vectorization provided in an embodiment of this disclosure is shown;

[0022] Figure 8 A flowchart illustrating a method for determining sign features according to an embodiment of this disclosure is shown;

[0023] Figure 9 A schematic diagram of the vectorized features of the detection sign provided in an embodiment of this disclosure is shown;

[0024] Figure 10A flowchart illustrating a method for determining the characteristics of a rod-shaped object according to an embodiment of this disclosure is shown;

[0025] Figure 11 A schematic diagram of the vectorized features of the detected rod-shaped object provided in an embodiment of this disclosure is shown;

[0026] Figure 12 A schematic diagram of the structure of a map feature determination device provided in an embodiment of this disclosure is shown;

[0027] Figure 13 A schematic block diagram of an example electronic device 1300 that can be used to implement embodiments of the present disclosure is shown. Detailed Implementation

[0028] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0029] ANP positioning layers provide basic map positioning services through lane lines, signs, and poles, supporting closed-loop ANP in highway and urban scenarios.

[0030] In related technologies, the ANP positioning layer only provides positioning information based on lane lines. Lane lines can only provide positioning constraints for lateral features and cannot determine the positioning of longitudinal features. This can lead to significant positioning errors for autonomous vehicles on roads. For example, when only lane lines are used for positioning, there are only lateral positioning constraints, making it impossible to determine longitudinal locations such as traffic lights or the distance to intersections.

[0031] Based on this, this disclosure provides a map feature determination method and apparatus. By detecting the position and attributes of vertical structural objects in an image, the method determines the position and three-dimensional geometric information of vertical structural objects in point cloud data. This yields the features of vertical structural objects in a high-precision map, providing localization constraints for the vertical features in the ANP positioning layer. This improves the recall rate of vertical structural object features in the high-precision map, further enhancing the visual positioning performance of autonomous vehicles.

[0032] The map feature determination method provided in this application can be applied to, for example... Figure 1In the application environment shown, terminal 101 communicates with server 102 via a network. Terminal 101 can be a camera device of an autonomous vehicle (i.e., an unmanned vehicle), such as a camera or radar. Terminal 101 is used to acquire images and point cloud data containing accurately structured longitudinal objects. Server 102 is used to detect longitudinal objects in the images and determine the features of the longitudinal objects. Terminal 101 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices, and server 102 can be implemented using a standalone server or a server cluster consisting of multiple servers.

[0033] It should be noted that the terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0034] The following embodiments will describe the map feature determination method and apparatus improved by this disclosure in conjunction with the accompanying drawings.

[0035] Figure 2 A flowchart illustrating a map feature determination method provided in an embodiment of this disclosure is shown, such as... Figure 2 As shown, the method may include:

[0036] In step S210, an image of the road is acquired, and point cloud data of the road is acquired.

[0037] In this embodiment of the disclosure, the device used to acquire road images can be an autonomous vehicle. The road images are acquired via camera devices included in the autonomous vehicle. These road images can be two-dimensional images. The two-dimensional images contain accurately structured longitudinal structural objects, which may include poles and / or signs. For example, the two-dimensional image can be a location map containing accurately structured geometrical position information of poles and / or signs.

[0038] In this disclosure, the point cloud data of the road can be a map used for assisted driving, wherein the map can be a high-definition map. High-definition maps can improve assisted driving information for autonomous vehicles. In this disclosure, the point cloud data of the high-definition map can be acquired through the vehicle's scanning device.

[0039] In step S220, vertical structural objects in the image are detected, and the detection location and attributes of the vertical structural objects are determined.

[0040] In this embodiment of the disclosure, a detection model provided by a perceived two-dimensional image is used to detect the detection location and attributes of a longitudinally structured object. These attributes can be used to characterize whether the longitudinally structured object is a rod-shaped object or a sign.

[0041] In step S230, features of the vertically structured object are determined in the point cloud data based on the detected location and attributes.

[0042] In this embodiment of the disclosure, based on the detection position and attributes of the vertical structural object, the position and three-dimensional geometric information of the vertical structural object can be determined in the point cloud data, thereby determining the features of the vertical structural object, and automatically labeling the features of the vertical structural object into the high-precision map.

[0043] The map feature determination method provided in this disclosure determines the position and attributes of vertically combined objects in an image by detecting such objects. This allows the use of point cloud data from a high-precision map to determine the features of the vertically combined objects, providing localization information of vertical features for autonomous vehicles, improving the recall rate of map features, and thus enhancing the visual localization effect of autonomous vehicles.

[0044] In this embodiment of the disclosure, rod-shaped objects in a longitudinal structural object can be detected; signs in a longitudinal structural object can also be detected.

[0045] In one embodiment of this disclosure, a detection model provided by a perceived two-dimensional image can be used to detect longitudinally structured objects. The detection model may include a rod-shaped object detection model and a sign detection model. The rod-shaped object detection model is used to detect rod-shaped objects within the longitudinally structured object, and the sign detection model is used to detect signs within the longitudinally structured object.

[0046] The detection model used in this disclosure can ensure consistency with the perception effect of autonomous vehicles, making the detection of longitudinal structural objects more accurate.

[0047] The following examples will illustrate how to detect vertical structural objects in an image and determine the detection location and attributes of the vertical structural objects.

[0048] Figure 3 This illustration shows a flowchart of a method for determining the detection position and attributes of a longitudinally structured object according to an embodiment of the present disclosure, as shown below. Figure 3 As shown, the method may include:

[0049] In step S310, a first detection box of the vertical structural object in the detected image is obtained, and a second detection box of the vertical structural object in the detected point cloud data is obtained.

[0050] In step S320, the detection position of the vertical structural object is determined in the point cloud data based on the overlapping area of ​​the first detection box and the second detection box.

[0051] In step S330, the properties of the longitudinal structural object are determined based on the detected position.

[0052] In this embodiment of the disclosure, vertical structural objects are detected in the image to obtain a first detection box containing the vertical structural object in the image; vertical structural objects are also detected in the point cloud data to obtain a second detection box containing the vertical structural object in the point cloud data.

[0053] The overlap area (IOU) of the first and second detection boxes is determined by projection, and the detection position and attributes of vertical structural objects in the point cloud data are determined based on the overlap area. The attributes can also be the three-dimensional geometric information of the vertical structural objects.

[0054] In one embodiment of this disclosure, if the overlapping area is greater than or equal to a preset area threshold, the position of the second detection box is determined as the detection position of the vertical structural object in the point cloud data.

[0055] In another embodiment of this disclosure, if the overlapping area is less than the area threshold, the first detection box of the vertical structural object in the image is re-determined until the overlapping area is greater than or equal to the area threshold, and the position of the second detection box is determined as the detection position of the vertical structural object in the point cloud data.

[0056] Figure 4 A flowchart illustrating a positioning layer generation method provided in an embodiment of this disclosure is shown. Figure 4 As shown, image boxes (i.e., first detection boxes) are detected for features in the original image, and features in the point cloud data are detected (second detection boxes). The overlapping area of ​​the detection boxes is determined by projection, and the calculated overlapping area determines whether to re-detect the detection boxes, thereby determining the position of the detection boxes. Point cloud data provides three-dimensional features of vertically structured objects; therefore, it is necessary to deduplicate the three-dimensional features to obtain the detection position of the vertically structured objects in the point cloud data.

[0057] The following examples illustrate how to determine the features of vertically structured objects in point cloud data based on detection location and attributes.

[0058] Figure 5 A flowchart illustrating a method for determining the features of a longitudinally structured object according to an embodiment of this disclosure is shown, such as... Figure 5 As shown, the method may include:

[0059] In step S510, the point cloud data is divided to obtain at least one sub-point cloud data, and labels are added to the sub-point cloud data.

[0060] In step S520, point cloud data corresponding to the vertical structure object is determined based on the label.

[0061] In step S530, based on the detection location and the attribute, target sub-point cloud data is determined in the point cloud data corresponding to the longitudinal structural object.

[0062] In step S540, local and global features of the vertical structure object are extracted based on the target sub-point cloud data.

[0063] In step S550, the features of the longitudinal structural object are determined based on local and global features.

[0064] In this embodiment of the disclosure, point cloud data can be divided to detect features of vertically structured objects in the divided point cloud data.

[0065] Furthermore, the point cloud data is divided to obtain at least one local point cloud data. For ease of description, this disclosure refers to the local point cloud data as sub-point cloud data. Each sub-point cloud data is then labeled using a semantic segmentation model. For example, if the sub-point cloud data represents a rod-shaped object, a rod-shaped object label is added; if the sub-point cloud data represents a sign, a sign label is added.

[0066] Based on the labels of each sub-point cloud data and the attributes of the currently detected vertical structural object, the target sub-point cloud data corresponding to the current vertical structural object is determined. Then, local and global features of the vertical structural object are extracted from the target sub-point cloud data. Based on the local and global features, projection verification is performed to determine the validity of the vertical structural object and to identify its features. These features are then labeled in a high-precision map; the labeled features are vectorized features. For example, Figure 6 A schematic diagram of the sign vectorization provided in an embodiment of this disclosure is shown. Figure 6 This disclosure presents a high-precision map after the vectorized features of the signboard are annotated. Figure 7 A schematic diagram of rod vectorization provided in an embodiment of this disclosure is shown. Figure 7 This disclosure presents a high-precision map after the vectorized features of the rod-shaped object are annotated.

[0067] In this disclosure, local features of a vertically structured object can be extracted based on target sub-point cloud data, and global features of the vertically structured object can be extracted based on the combined target sub-point cloud data.

[0068] In one embodiment of this disclosure, if a sign is detected within a vertically oriented structural object, the point cloud data can be divided into block-shaped point cloud data to obtain at least one submap. The sign can be a triangular sign or a rectangular sign, etc.

[0069] The following examples will illustrate the implementation of high-precision map point cloud data detection signs.

[0070] Figure 8 A flowchart illustrating a method for determining sign features according to an embodiment of this disclosure is shown, such as... Figure 8 As shown, the method may include:

[0071] In step S810, based on the detection location, a first candidate sub-point cloud data block is determined in the sub-point cloud data block labeled as a sign tag.

[0072] In step S820, the first candidate sub-point cloud data block is filtered based on the attributes to obtain the second candidate sub-point cloud data block.

[0073] In step S830, the second candidate sub-point cloud data block is filtered based on the size threshold of the sign and the confidence level of the second candidate sub-point cloud data block to determine the target sub-point cloud data block.

[0074] In this embodiment of the disclosure, a sub-point cloud data block labeled as a sign label is determined based on the label of each sub-point cloud data. Further, sign sub-point cloud data blocks within a certain range are extracted based on the detection location. A first candidate sub-point cloud data block is determined from the sign sub-point cloud data.

[0075] Noise is filtered out using a threshold method on the signboard, and facade data of the first candidate sub-point cloud data block is extracted based on Singular Value Decomposition (SVD). Then, each first candidate sub-point cloud data block is divided into second candidate sub-point cloud data blocks using distance clustering.

[0076] The size and confidence level of each second candidate sub-point cloud data block are determined, and the confidence levels are sorted. Second candidate sub-point cloud data blocks are filtered based on a sign size threshold, and then further filtered according to the confidence level sequence to determine the target sub-point cloud data block.

[0077] Figure 9 A schematic diagram of the vectorized features of the detection sign provided in an embodiment of this disclosure is shown, such as... Figure 9As shown, the sensor detects the image and triangulates the detected sign. Triangulation can be understood as recovering the 3D coordinates of feature points based on their projections across multiple cameras. This high-precision triangulation... Figure 3 The point cloud data is divided into sub-maps. Multiple sub-maps can be merged to form a larger block. Each point cloud is labeled with semantic tags (i.e., category labels) using a semantic segmentation model, and the detection location is reconstructed from the image bounding box. Local features of the sign are extracted in the submap, and global sign features are extracted in the block. Specifically, the semantic point cloud is preprocessed to extract sign point clouds within a certain range, outputting semantic candidate signs. Noise is filtered by thresholding, and then the facade of the sign is extracted using SVD. The signs are then divided into candidate signs by distance clustering. These candidate signs are further filtered by size and ranked by point confidence, finally obtaining the vectorized features of the target sign.

[0078] In another embodiment of this disclosure, if the detected vertical structure includes rod-shaped objects, the point cloud data can be divided into point cloud data layers in the vertical direction to obtain at least one sub-point cloud data layer.

[0079] The following examples will illustrate the implementation of detecting rod-shaped objects using high-precision map point cloud data.

[0080] Figure 10 A flowchart illustrating a method for determining the characteristics of a rod-shaped object according to an embodiment of this disclosure is shown, as follows. Figure 10 As shown, the method may include:

[0081] In step S1010, based on the detection location, the sub-point cloud data layers to be merged are determined in the sub-point cloud data layers with rod-shaped labels.

[0082] In step S1020, candidate sub-point cloud data layers are determined by region growing in the sub-point cloud data layers to be merged, based on the properties of the rod-like objects.

[0083] In step S1030, the target sub-point cloud data layer is determined from the candidate sub-point cloud data layers based on the size of the rod and the confidence level of the candidate sub-point cloud data layers.

[0084] In this embodiment, within a sub-point cloud data layer labeled with rod-shaped tags, seed points are used to locate sub-point cloud data layers to be merged. Candidate sub-point cloud data layers are determined through region growing within the sub-point cloud data layers to be merged. The size and confidence level of each candidate sub-point cloud data layer are determined. The candidate sub-point cloud data layers are filtered by the size of the rod-shaped tags, and within the size-filtered candidate sub-point cloud data layers, further filtering is performed based on the confidence level to obtain the target sub-point cloud data layer.

[0085] Figure 11 A schematic diagram of the vectorized features of the detected rod-shaped object provided in an embodiment of this disclosure is shown, such as... Figure 11 As shown, the sensor detects the image and triangulates the detected rod-shaped objects. First, the semantic point cloud is segmented vertically. Then, the point cloud is merged by finding seed points. Candidate rods are extracted using a region growing method. Finally, these candidate rods are filtered by size and sorted by point count confidence to obtain the target rod-shaped object.

[0086] Based on and Figure 2 The method shown follows the same principle. Figure 12 A schematic diagram of a map feature determination device provided in an embodiment of this disclosure is shown, such as... Figure 12 As shown, the map feature determination device 1200 may include:

[0087] The acquisition module 1201 is used to acquire an image of the road and acquire point cloud data of the road; the determination module 1202 is used to detect vertical structural objects in the image and determine the detection position and attributes of the vertical structural objects; the determination module 1202 is also used to determine the features of the vertical structural objects in the point cloud data based on the detection position and attributes.

[0088] In this embodiment of the disclosure, the determining module 1202 is used to detect rod-shaped objects in the longitudinal structural object and to detect signs in the longitudinal structural object.

[0089] In this embodiment of the disclosure, the determining module 1202 is used to obtain a first detection box for detecting vertical structural objects in the image, and to obtain a second detection box for detecting vertical structural objects in the point cloud data; based on the overlapping area of ​​the first detection box and the second detection box, to determine the detection position of the vertical structural object in the point cloud data; and to determine the attributes of the vertical structural object based on the detection position.

[0090] In this embodiment of the disclosure, the determining module 1202 is used to determine the position of the second detection box as the detection position of the vertical structural object in the point cloud data if the overlapping area is greater than or equal to a preset area threshold; or, if the overlapping area is less than the area threshold, to redetermine the first detection box of the vertical structural object in the image until the overlapping area is greater than or equal to the area threshold.

[0091] In this embodiment of the disclosure, the determining module 1202 is used to divide the point cloud data to obtain at least one sub-point cloud data and add labels to the sub-point cloud data; based on the labels, determine the point cloud data corresponding to the vertical structural object; based on the detection position and the attribute, determine the target sub-point cloud data in the point cloud data corresponding to the vertical structural object; based on the target sub-point cloud data, extract the local features and global features of the vertical structural object; and based on the local features and the global features, determine the features of the vertical structural object.

[0092] In this embodiment of the disclosure, the determining module 1202 is used to extract local features of the vertical structure object based on the target sub-point cloud data; and to combine the target sub-point cloud data and extract global features of the vertical structure object based on the combined target sub-point cloud data.

[0093] In this embodiment of the disclosure, the vertical structural object is a signboard, and the point cloud data is divided to obtain at least one sub-point cloud data block;

[0094] The determining module 1202 is used to determine a first candidate sub-point cloud data block in the sub-point cloud data block labeled as a sign based on the detection location; filter the first candidate sub-point cloud data block based on the attribute to obtain a second candidate sub-point cloud data block; filter the second candidate sub-point cloud data block based on the size threshold of the sign and the confidence level of the second candidate sub-point cloud data block to determine the target sub-point cloud data block.

[0095] In this embodiment of the disclosure, the vertical structure is a rod-shaped object, and the point cloud data is divided to obtain at least one sub-point cloud data layer;

[0096] The determining module 1202 is used to determine, based on the detection location, a sub-point cloud data layer to be merged in the sub-point cloud data layer labeled with a rod-shaped object; to determine a candidate sub-point cloud data layer in the sub-point cloud data layer to be merged by region growing based on the attributes of the rod-shaped object; and to determine a target sub-point cloud data layer in the candidate sub-point cloud data layer based on the size of the rod-shaped object and the confidence level of the candidate sub-point cloud data layer.

[0097] The acquisition, storage, and application of user personal information involved in the technical solution disclosed herein comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0098] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0099] Figure 13 A schematic block diagram of an example electronic device 1300 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0100] like Figure 13 As shown, device 1300 includes a computing unit 1301, which can perform various appropriate actions and processes according to a computer program stored in read-only memory (ROM) 1302 or a computer program loaded from storage unit 1308 into random access memory (RAM) 1303. The RAM 1303 may also store various programs and data required for the operation of device 1300. The computing unit 1301, ROM 1302, and RAM 1303 are interconnected via bus 1304. Input / output (I / O) interface 1305 is also connected to bus 1304.

[0101] Multiple components in device 1300 are connected to I / O interface 1305, including: input unit 1306, such as keyboard, mouse, etc.; output unit 1307, such as various types of monitors, speakers, etc.; storage unit 1308, such as disk, optical disk, etc.; and communication unit 1309, such as network card, modem, wireless transceiver, etc. Communication unit 1309 allows device 1300 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0102] The computing unit 1301 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 1301 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 1301 performs the various methods and processes described above, such as map feature vectorization methods. For example, in some embodiments, the map feature vectorization method may be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 1308. In some embodiments, part or all of the computer program may be loaded and / or installed on device 1300 via ROM 1302 and / or communication unit 1309. When the computer program is loaded into RAM 1303 and executed by the computing unit 1301, one or more steps of the map feature vectorization method described above may be performed. Alternatively, in other embodiments, the computing unit 1301 may be configured to perform a map feature vectorization method by any other suitable means (e.g., by means of firmware).

[0103] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0104] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0105] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0106] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0107] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with embodiments of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0108] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0109] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0110] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for determining map features, the method comprising: Acquire image data of the road and acquire point cloud data of the road; Detect vertical structural objects in the image, and determine the detection location and attributes of the vertical structural objects; Based on the detection location and the attribute, the features of the vertical structure object are determined in the point cloud data, and the features of the vertical structure object are labeled in the high-precision map. The determination of the detection position and attributes of the longitudinal structural object includes: Obtain a first detection box for detecting vertical structural objects in the image, and obtain a second detection box for detecting vertical structural objects in the point cloud data; Based on the overlapping area of ​​the first detection box and the second detection box, the detection position of the vertical structure object is determined in the point cloud data; The properties of the longitudinal structural object are determined based on the detection location.

2. The method according to claim 1, wherein, The detection of vertically structured objects in the image includes: Detect rod-shaped objects in the longitudinal structural object; detect signs in the longitudinal structural object.

3. The method according to claim 1, wherein, Determining the detection position of the vertical structural object in the point cloud data based on the overlapping area of ​​the first and second detection boxes includes: If the overlapping area is greater than or equal to a preset area threshold, the position of the second detection box is determined as the detection position of the vertical structural object in the point cloud data. or If the overlapping area is less than the area threshold, the first detection box of the vertical structural object in the image is redefined until the overlapping area is greater than or equal to the area threshold.

4. The method according to claim 1, wherein, The step of determining the features of the vertical structure object in the point cloud data based on the detected location and attributes includes: The point cloud data is divided to obtain at least one sub-point cloud data, and labels are added to the sub-point cloud data. Based on the label, the point cloud data corresponding to the vertical structure object is determined; Based on the detection location and the attribute, target sub-point cloud data is determined in the point cloud data corresponding to the vertical structural object; Based on the target sub-point cloud data, extract the local and global features of the vertical structure object; Based on the local features and the global features, the features of the vertical structure object are determined.

5. The method according to claim 4, wherein, The step of extracting local and global features of the vertical structure object based on the target sub-point cloud data includes: Based on the target sub-point cloud data, local features of the vertical structure object are extracted; and The target sub-point cloud data are combined, and based on the combined target sub-point cloud data, the global features of the vertical structure object are extracted.

6. The method according to claim 4, wherein, The vertical structural object includes a signboard, and the point cloud data is divided to obtain at least one sub-point cloud data block; The step of determining target sub-point cloud data from the point cloud data corresponding to the longitudinal structural object based on the detection location and the attribute includes: Based on the detection location, a first candidate sub-point cloud data block is determined in the sub-point cloud data block whose label is a signboard label; Based on the aforementioned attributes, the first candidate sub-point cloud data block is filtered to obtain the second candidate sub-point cloud data block. Based on the size threshold of the sign and the confidence level of the second candidate sub-point cloud data block, the second candidate sub-point cloud data block is filtered to determine the target sub-point cloud data block.

7. The method according to claim 4, wherein, The longitudinal structural object includes rod-shaped objects, and the point cloud data is divided to obtain at least one sub-point cloud data layer; The step of determining target sub-point cloud data from the point cloud data corresponding to the longitudinal structural object based on the detection location and the attribute includes: Based on the detection location, in the sub-point cloud data layer with the label being a rod-shaped label, the sub-point cloud data layer to be merged is determined; Based on the properties of the rod-shaped object, candidate sub-point cloud data layers are determined in the sub-point cloud data layers to be merged by region growing. Based on the size of the rod and the confidence level of the candidate sub-point cloud data layers, the target sub-point cloud data layer is determined from the candidate sub-point cloud data layers.

8. A map feature determination device, the device comprising: The acquisition module is used to acquire images of the road and acquire point cloud data of the road; A determination module is used to detect vertical structural objects in the image and determine the detection location and attributes of the vertical structural objects; The determining module is further configured to determine the features of the vertical structure object in the point cloud data based on the detection location and the attribute, and to annotate the features of the vertical structure object in the high-precision map; The determining module is used to detect vertical structural objects in the image and determine the detection location and attributes of the vertical structural objects, including: Obtain a first detection box for detecting vertical structural objects in the image, and obtain a second detection box for detecting vertical structural objects in the point cloud data; Based on the overlapping area of ​​the first detection box and the second detection box, the detection position of the vertical structure object is determined in the point cloud data; The properties of the longitudinal structural object are determined based on the detection location.

9. The apparatus according to claim 8, wherein, The determining module is used for: Detect rod-shaped objects in the longitudinal structural object; detect signs in the longitudinal structural object.

10. The apparatus according to claim 8, wherein, The determining module is used for: If the overlapping area is greater than or equal to a preset area threshold, the position of the second detection box is determined as the detection position of the vertical structural object in the point cloud data. or If the overlapping area is less than the area threshold, the first detection box of the vertical structural object in the image is redefined until the overlapping area is greater than or equal to the area threshold.

11. The apparatus according to claim 8, wherein, The determining module is used for: The point cloud data is divided to obtain at least one sub-point cloud data, and labels are added to the sub-point cloud data. Based on the label, the point cloud data corresponding to the vertical structure object is determined; Based on the detection location and the attribute, target sub-point cloud data is determined in the point cloud data corresponding to the vertical structural object; Based on the target sub-point cloud data, extract the local and global features of the vertical structure object; Based on the local features and the global features, the features of the vertical structure object are determined.

12. The apparatus according to claim 11, wherein, The determining module is used for: Based on the target sub-point cloud data, local features of the vertical structure object are extracted; and The target sub-point cloud data are combined, and based on the combined target sub-point cloud data, the global features of the vertical structure object are extracted.

13. The apparatus according to claim 11, wherein, The vertical structural object includes a signboard, and the point cloud data is divided to obtain at least one sub-point cloud data block; The determining module is used for: Based on the detection location, a first candidate sub-point cloud data block is determined in the sub-point cloud data block whose label is a signboard label; Based on the aforementioned attributes, the first candidate sub-point cloud data block is filtered to obtain the second candidate sub-point cloud data block. Based on the size threshold of the sign and the confidence level of the second candidate sub-point cloud data block, the second candidate sub-point cloud data block is filtered to determine the target sub-point cloud data block.

14. The apparatus according to claim 11, wherein, The longitudinal structural object includes rod-shaped objects, and the point cloud data is divided to obtain at least one sub-point cloud data layer; The determining module is used for: Based on the detection location, in the sub-point cloud data layer with the label being a rod-shaped label, the sub-point cloud data layer to be merged is determined; Based on the properties of the rod-shaped object, candidate sub-point cloud data layers are determined in the sub-point cloud data layers to be merged by region growing. Based on the size of the rod and the confidence level of the candidate sub-point cloud data layers, the target sub-point cloud data layer is determined from the candidate sub-point cloud data layers.

15. An electronic device comprising: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-7.

16. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-7.

17. A computer program product comprising a computer program that, when executed by a processor, implements the method according to any one of claims 1-7.

Images