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

Abstract

The application discloses a high-precision map construction method and device, electronic equipment and a storage medium. The method comprises the following steps: obtaining a road element semantic segmentation result through image information collected by a camera; correcting a pose transformation relationship of a vehicle relative to a ground coordinate system according to vehicle positioning information, wherein the vehicle positioning information comprises positioning information output by a combined navigation device; determining a pose transformation relationship of the camera relative to the ground coordinate system according to the pose transformation relationship of the vehicle relative to the ground coordinate system; projecting a ground grid point into the road element semantic segmentation result to obtain a semantic value of each ground grid point according to the pose transformation relationship of the camera relative to the ground coordinate system, wherein the ground grid point is obtained by dividing the ground coordinate system at a self-defined resolution; and constructing the high-precision map according to the semantic value of each ground grid point. Through the application, a high-precision and timely-updated high-precision map can be constructed.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automatic driving, in particular to a high-precision map construction method and device, electronic equipment and a storage medium. BACKGROUND

[0002] In the field of automatic driving, a high-precision map is an important component of high-level automatic driving and is the basis for vehicle high-precision positioning and navigation planning. A high-precision map is usually constructed by using a laser radar scan or a multi-camera to collect and construct a high-precision semantic map.

[0003] In related technologies, a laser radar is expensive and is not conducive to large-scale deployment and timely updating of a map. Although a multi-camera reduces the cost and includes semantic types, the accuracy is difficult to guarantee. SUMMARY

[0004] Embodiments of the present application provide a high-precision map construction method and device, electronic equipment and a storage medium to provide a high-precision map that can be updated in a timely manner.

[0005] Embodiments of the present application use the following technical solutions:

[0006] In a first aspect, embodiments of the present application provide a high-precision map construction method, and the method comprises:

[0007] Obtaining a road element semantic segmentation result from image information collected by a plurality of cameras;

[0008] Correcting a pose transformation relationship of a vehicle relative to a ground coordinate system according to vehicle positioning information, wherein the vehicle positioning information comprises positioning information output by a combined navigation device;

[0009] Determining a pose transformation relationship of a camera relative to the ground coordinate system according to the pose transformation relationship of the vehicle relative to the ground coordinate system;

[0010] Projecting a ground grid point into the road element semantic segmentation result to obtain a semantic value of each ground grid point according to the pose transformation relationship of the camera relative to the ground coordinate system, wherein the ground grid point is obtained by dividing the ground coordinate system at a self-defined resolution;

[0011] Constructing the high-precision map according to the semantic value of each ground grid point.

[0012] In a second aspect, embodiments of the present application also provide a high-precision map construction device, and the device comprises:

[0013] A semantic segmentation module configured to obtain a road element semantic segmentation result from image information collected by a plurality of cameras;

[0014] The correction module is used to correct the pose transformation relationship of the vehicle relative to the ground coordinate system based on the vehicle positioning information, wherein the vehicle positioning information includes the positioning information output by the integrated navigation device.

[0015] The determination module is used to determine the pose transformation relationship of the camera relative to the ground coordinate system based on the pose transformation relationship of the vehicle relative to the ground coordinate system.

[0016] The projection module is used to project ground grid points onto the semantic segmentation result of the road element according to the pose transformation relationship of the camera relative to the ground coordinate system to obtain the semantic value of each ground grid point, wherein the ground grid points are obtained by dividing in the ground coordinate system with a custom resolution;

[0017] A construction module is used to construct the high-precision map based on the semantic values ​​of each ground grid point.

[0018] Thirdly, embodiments of this application also provide an electronic device, including: a processor; and a memory arranged to store computer-executable instructions, which, when executed, cause the processor to perform the above-described method.

[0019] Fourthly, embodiments of this application also provide a computer-readable storage medium that stores one or more programs, which, when executed by an electronic device including multiple applications, cause the electronic device to perform the above-described method.

[0020] The above-described technical solutions adopted in the embodiments of this application can achieve the following beneficial effects:

[0021] First, semantic segmentation results of road elements are obtained from image information acquired by multiple cameras. Then, the pose transformation relationship of the vehicle relative to the ground coordinate system is corrected based on vehicle positioning information, and the pose transformation relationship of the camera relative to the ground coordinate system is determined based on the vehicle pose transformation relationship relative to the ground coordinate system. By using vehicle positioning information to correct vehicle body undulation in real time, the accurate pose transformation relationship from the camera to the ground can be obtained, thereby improving mapping accuracy. Next, based on the pose transformation relationship of the camera relative to the ground coordinate system, ground grid points are projected onto the semantic segmentation results of the road elements to obtain the semantic value of each ground grid point, making each ground grid point evenly distributed. Finally, the high-precision map is constructed based on the semantic values ​​of each ground grid point, ensuring timely data updates. Attached Figure Description

[0022] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0023] Figure 1 FIG. 1 is a flowchart of a high-definition map construction method according to an embodiment of the present application.

[0024] Figure 2 FIG. 2 is a schematic diagram of a high-definition map construction device according to an embodiment of the present application.

[0025] Figure 3 FIG. 3 is a schematic diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0026] In order to make the objectives, technical solutions, and advantages of the present application clearer, the technical solutions of the present application will be described below in detail with reference to the embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0027] The inventors found in the related art that when a high-definition map is constructed by high-density scanning using a laser radar, the position accuracy of the point cloud is high, but the semantic information is inaccurate or even lacking, and the construction of the map needs to rely on expensive equipment, which is not conducive to large-scale deployment and timely updating of the map.

[0028] In addition, if a high-definition semantic map is constructed by using a camera to collect, although the cost can be reduced and the semantic type can be included, the construction process is seriously affected by the ups and downs of the vehicle body (road), and the accuracy is difficult to guarantee.

[0029] In view of the above problems, the high-definition map construction method according to an embodiment of the present application uses image semantic segmentation and high-precision combined navigation positioning to construct at least one road element map, thereby providing a lane-level high-definition map for automatic driving positioning.

[0030] The technical solutions provided by the embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0031] The present application provides a high-definition map construction method, as shown in Figure 1 The present application provides a high-definition map construction method, as shown in

[0032] Step S110: obtaining a road element semantic segmentation result from image information collected by a camera.

[0033] The image information is acquired by the camera, and the image information is subjected to road element semantic segmentation to obtain a road element semantic segmentation result. The acquired image information can be subjected to semantic segmentation by using various semantic segmentation models to obtain the road element semantic segmentation result, and the specific selection of the semantic segmentation model is not limited in the present application, and can be selected by the person skilled in the art according to the actual situation.

[0034] It should be noted that the extrinsic parameters of the plurality of cameras are calibrated in advance, so that the high-precision semantic map acquired and constructed by the plurality of cameras contains semantic types.

[0035] It can be understood that the road element semantics are mainly road elements such as lane lines, arrows, stop lines, sidewalks and the like related to high-precision map construction, corresponding to the road element semantics.

[0036] In step S120, the pose transformation relationship of the vehicle relative to the ground coordinate system is corrected according to vehicle positioning information, wherein the vehicle positioning information includes positioning information output by a combined navigation device.

[0037] The vehicle positioning information includes positioning information output by a combined navigation device, and the pose transformation relationship of the vehicle relative to the ground coordinate system is corrected by vehicle information obtained by combined navigation. Here, the problem of vehicle body heave caused by uneven road surface is mainly considered, and the vehicle body heave problem is corrected in real time by using vehicle positioning information, and the accurate pose of the camera to the ground is further obtained, that is, the vehicle-to-ground pose is corrected by using the vehicle positioning information.

[0038] At the same time, the correction process using the vehicle positioning information is performed in real time, which ensures the timeliness and accuracy of map construction.

[0039] In some embodiments, the roll angle and yaw angle errors of the camera to the ground caused by the vehicle body heave can be corrected in real time.

[0040] In some embodiments, the road elements in the vehicle body coordinate system can be spliced according to the combined navigation positioning information, and real-time correction can be realized after splicing.

[0041] In step S130, the pose transformation relationship of the camera relative to the ground coordinate system is determined according to the pose transformation relationship of the vehicle relative to the ground coordinate system.

[0042] The pose transformation relationship of the camera relative to the ground coordinate system can be further determined by the pose transformation relationship of the vehicle relative to the ground coordinate system after correction and the extrinsic parameters of the camera. That is, the pose of the camera to the ground is corrected by extrinsic parameter conversion.

[0043] Step S140: Based on the pose transformation relationship of the camera relative to the ground coordinate system, the ground grid points are projected onto the semantic segmentation result of the road element to obtain the semantic value of each ground grid point, wherein the ground grid points are obtained by dividing the ground coordinate system with a custom resolution.

[0044] By utilizing the pose transformation relationship of the camera relative to the ground coordinate system, the ground grid points are projected onto the semantic segmentation image to obtain the semantic value of each ground grid point.

[0045] Step S150: Construct the high-precision map based on the semantic value of each ground grid point.

[0046] Since it includes multiple ground grid points, the high-precision map is constructed based on the semantic value of each ground grid point (e.g., belonging to a road semantic element). That is, if the semantic value of a ground grid point belongs to a road element, the ground grid point is rotated to the initial vehicle coordinate system and then added to the map to complete the mapping function.

[0047] High-precision maps are a crucial component of advanced autonomous driving, serving as the foundation for high-precision vehicle positioning and navigation planning. It is understood that the high-precision map construction method in this application is designed for scenarios requiring local or real-time updates to high-precision maps, achieving both high accuracy and timely updates.

[0048] The high-precision map construction method described above uses image semantic segmentation results and perspective transformation algorithms to transfer road surface elements in the image to the vehicle coordinate system. Then, based on the integrated navigation positioning, the road surface elements in the vehicle coordinate system are stitched together, and the roll angle error and pitch angle error caused by vehicle body undulation are corrected in real time.

[0049] It should be noted that multiple coordinate systems need to be established before implementing the above scheme. The definitions of multiple coordinate systems are as follows:

[0050] Vehicle coordinate system (rear axle center as origin, front left top -> xyz).

[0051] Combined navigation coordinate system (GPS / RTK+IMU, front left top -> xyz).

[0052] Camera coordinate system (bottom right front -> xyz).

[0053] Ground coordinate system (the projection of the rear axle center of the vehicle body onto the ground is the origin of the coordinate system, front left top -> xyz).

[0054] In an embodiment of the present application, the plurality of cameras and the combined navigation device are pre-installed on the vehicle, and the method further comprises: simultaneously receiving image information collected by the plurality of cameras and vehicle positioning information output by the combined navigation device, wherein the timestamps of the camera image information and the vehicle positioning information are aligned; performing semantic segmentation of road elements on the image information, and caching the vehicle positioning information to a preset processing queue, wherein the combined navigation device and the plurality of cameras are pre-calibrated with respect to the vehicle body.

[0055] In specific implementation, the image information collected by the plurality of cameras and the vehicle positioning information output by the combined navigation device are simultaneously received, and then the image information is subjected to semantic segmentation of road elements, and the vehicle positioning information is processed and cached to a preset processing queue.

[0056] It should be noted that the timestamps of the camera image information and the vehicle positioning information are aligned, or for timestamps that are not aligned, the timestamps are aligned first and then processed.

[0057] In addition, the combined navigation device and the plurality of cameras are pre-calibrated with respect to the vehicle body. It can be understood that the combined navigation device includes but is not limited to an IMU+RTK / GPS combined navigation device.

[0058] In an embodiment of the present application, the vehicle pose transformation relationship with respect to the ground coordinate system is corrected according to the vehicle positioning information, wherein the vehicle positioning information includes positioning information output by the combined navigation device, and the method comprises: calculating a relative displacement and a height change amount corresponding to the vehicle positioning information at the head and tail of the preset processing queue; in the case that the relative displacement is greater than a preset threshold, calculating an angle between the ground and the ground plane; calculating a pitch angle of the vehicle with respect to the ground according to the pitch angle of the vehicle positioning and the angle between the ground and the ground plane; and obtaining the vehicle pose transformation relationship with respect to the ground coordinate system according to the pitch angle of the vehicle with respect to the ground and the roll angle of the vehicle positioning.

[0059] In specific implementation, the relative displacement ds and the height change amount dh of the vehicle positioning at the head and tail of the queue in the preset processing queue are calculated.

[0060] Further, when the relative displacement ds is greater than the threshold, the angle theta between the ground and the ground plane is calculated as theta=arcsin(dh / dx), and the pitch angle of the vehicle with respect to the ground is calculated using the pitch angle of the vehicle positioning and the angle theta between the ground and the ground plane.

[0061]

[0062] wherein dh is the height variation, and theta is the angle between the ground and the horizontal plane.

[0063] The relative pose transformation of the vehicle relative to the ground coordinate system is also needed according to the pitch angle of the vehicle relative to the ground and the roll angle of the vehicle positioning.

[0064] The processing mode of the relative displacement of the front and rear queue heads and the tail of the preset processing queue and the height variation is adopted, the angle between the ground and the horizontal plane can be calculated in the case that the relative displacement is greater than a preset threshold, and the pitch angle of the vehicle relative to the ground is calculated according to the pitch angle of the vehicle positioning and the angle between the ground and the horizontal plane. Thus, the problem of body heave is eliminated and optimized.

[0065] In an embodiment of the present application, the vehicle positioning information output by the combined navigation device includes the rotation of the vehicle body coordinate system relative to the northeast sky coordinate system, the planar position in the UTM (Universal Transverse Mercator Grid System) coordinate system, and the altitude, the determination of the pose transformation relationship of the camera relative to the ground coordinate system according to the pose transformation relationship of the vehicle relative to the ground coordinate system includes: obtaining the pose transformation relationship of the vehicle relative to the ground coordinate system according to the pitch angle of the vehicle relative to the ground and the roll angle of the vehicle positioning, correcting the roll angle and pitch angle errors of the camera to the ground caused by body heave in real time, and determining the pose transformation relationship of the camera relative to the ground coordinate system.

[0066] In specific implementation, the pose transformation relationship of the vehicle relative to the ground coordinate system is obtained according to the pitch angle of the vehicle relative to the ground and the roll angle of the vehicle positioning. The roll angle and pitch angle errors of the camera to the ground caused by body heave are corrected in real time, and the pose transformation relationship of the camera relative to the ground coordinate system is determined according to the corrected pose transformation relationship of the vehicle relative to the ground coordinate system.

[0067] In an embodiment of the present application, the construction of the high-definition map according to the semantic value of each ground grid point includes: if the semantic value of the ground grid point belongs to a road element, the corresponding ground grid point is rotated to the vehicle body coordinate system and then updated into the high-definition map, wherein the road element at least includes one of the following: lane line, arrow, stop line, and sidewalk.

[0068] In specific implementation, the lane line, arrow, stop line, sidewalk and other road elements need to be updated into the high-definition map in time, so the ground grid points corresponding to the road elements according to the semantic values are rotated to the vehicle coordinate system, so as to be updated into the high-definition map.

[0069] In an embodiment of the present application, the determining of the pose transformation relationship of the camera relative to the ground coordinate system according to the pose transformation relationship of the vehicle relative to the ground coordinate system comprises: determining the pose transformation relationship of the camera relative to the ground coordinate system according to the pose transformation relationship of the vehicle relative to the ground coordinate system and the extrinsic parameters of the plurality of cameras to the vehicle.

[0070] In specific implementation, after the relative pose transformation relationship of the vehicle relative to the ground coordinate system is corrected, the pose transformation relationship of the camera to the ground coordinate system is obtained through the camera extrinsic parameters.

[0071] In an embodiment of the present application, the ground grid points are obtained by the following method: a region of interest is determined according to a self-defined horizontal length and horizontal width in the ground coordinate system; and the region of interest is divided into grids at a self-defined resolution to obtain the ground grid points, and the horizontal height value is 0.

[0072] In specific implementation, before the semantic segmentation is performed, a region of interest with a self-defined length and width is preset in the ground coordinate system, the region is divided into grids at a self-defined resolution, and the grid point coordinates are in the form of (x, y, 0). The use of the grid points divided in the ground coordinate system can achieve a self-defined resolution, and compared with the direct projection of the image pixel points to the ground in the related art, the obtained map points are more uniformly distributed.

[0073] The present application also provides a high-definition map construction device 200, as shown in Figure 2 A structure diagram of the high-definition map construction device in the embodiment of the present application is provided, and the high-definition map construction device 200 at least comprises a semantic segmentation module 210, a correction module 220, a determination module 230, a projection module 240 and a construction module 250, wherein:

[0074] In an embodiment of the present application, the semantic segmentation module 210 is specifically configured to obtain a road element semantic segmentation result through image information collected by the plurality of cameras.

[0075] The image information is collected through the plurality of cameras, and the road element semantic segmentation is performed on the image information to obtain a road element semantic segmentation result. The collected image information can be segmented by using a plurality of semantic segmentation models to obtain a road element semantic segmentation result, and the specific selection of the semantic segmentation model is not limited in the present application, and can be selected by the person skilled in the art according to the actual situation.

[0076] It should be noted that the extrinsic parameters of the plurality of cameras are pre-calibrated, so that a high-precision semantic map is collected and constructed by using the plurality of cameras, and the high-precision semantic map contains semantic types.

[0077] It can be understood that the road element semantics are corresponding road element semantics of road elements such as lane lines, arrows, stop lines, sidewalks and the like, which are mainly related to high-precision map construction.

[0078] In an embodiment of the present application, the correction module 220 is specifically configured to correct a pose transformation relationship of the vehicle relative to the ground coordinate system according to vehicle positioning information, wherein the vehicle positioning information includes positioning information output by a combined navigation device.

[0079] The vehicle positioning information includes positioning information output by a combined navigation device, and the vehicle information obtained by the combined navigation is used to correct the pose transformation relationship of the vehicle relative to the ground coordinate system. Here, the vehicle positioning information is used to correct the body heave problem in real time, and the accurate pose of the camera to the ground can be further obtained. That is, the vehicle positioning information is used to correct the pose of the vehicle to the ground.

[0080] Meanwhile, the correction process using the vehicle positioning information is performed in real time, thereby ensuring the timeliness and accuracy of map construction.

[0081] In some embodiments, the roll angle and yaw angle errors of the camera to the ground caused by the body heave can be corrected in real time.

[0082] In some embodiments, the road element under the vehicle coordinate system can be spliced according to the combined navigation positioning information, and real-time correction can be realized after splicing.

[0083] In an embodiment of the present application, the determination module 230 is specifically configured to determine a pose transformation relationship of the camera relative to the ground coordinate system according to the pose transformation relationship of the vehicle relative to the ground coordinate system.

[0084] Through the pose transformation relationship of the vehicle relative to the ground coordinate system after correction and the extrinsic parameters of the camera, the pose transformation relationship of the camera relative to the ground coordinate system can be further determined. That is, the pose of the camera to the ground is corrected through extrinsic parameter conversion.

[0085] In an embodiment of the present application, the projection module 240 is specifically configured to project a ground grid point into the road element semantic segmentation result to obtain a semantic value of each ground grid point according to the pose transformation relationship of the camera relative to the ground coordinate system, wherein the ground grid point is obtained by dividing the ground coordinate system at a self-defined resolution.

[0086] By using the pose transformation relationship of the camera relative to the ground coordinate system, the ground grid points are projected into the semantic segmentation image to obtain the semantic value of each ground grid point.

[0087] In an embodiment of the present application, the construction module 250 is specifically configured to construct the high-definition map according to the semantic value of each ground grid point.

[0088] Since a plurality of ground grid points are included, the high-definition map is constructed according to the semantic value of each ground grid point, such as belonging to a road semantic element. That is, if the semantic value of the ground grid point belongs to a road element, the ground grid point is rotated to the initial vehicle coordinate system and then added to the map to complete the mapping function.

[0089] It can be understood that the high-definition map construction device described above can realize each step of the high-definition map construction method provided in the foregoing embodiments, and the related explanations about the high-definition map construction method are all applicable to the high-definition map construction device, which will not be repeated here.

[0090] Figure 3 is a structural schematic diagram of an electronic device of an embodiment of the present application. Please refer to Figure 3 At the hardware level, the electronic device includes a processor, and optionally further includes an internal bus, a network interface, and a memory. The memory can include a memory such as a high-speed random-access memory (RAM), and can also include a non-volatile memory such as at least one disk memory. Of course, the electronic device can also include other hardware required by the business.

[0091] The processor, network interface, and memory can be connected to each other through an internal bus, which can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 3 In the figure, only one bidirectional arrow is used to represent the bus, but it does not mean that there is only one bus or only one type of bus.

[0092] The memory is used to store programs. Specifically, the program can include program code, and the program code includes computer operation instructions. The memory can include a memory and a non-volatile memory, and provides instructions and data to the processor.

[0093] The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs, and forms a high-definition map construction device at a logical level. The processor executes the program stored in the memory, and is specifically configured to perform the following operations:

[0094] obtain a road element semantic segmentation result from image information collected by the plurality of cameras;

[0095] correct a pose transformation relationship of the vehicle relative to the ground coordinate system according to vehicle positioning information, wherein the vehicle positioning information comprises positioning information output by a combined navigation device;

[0096] determine a pose transformation relationship of the camera relative to the ground coordinate system according to the pose transformation relationship of the vehicle relative to the ground coordinate system;

[0097] project a ground grid point into the road element semantic segmentation result to obtain a semantic value of each ground grid point according to the pose transformation relationship of the camera relative to the ground coordinate system, wherein the ground grid point is obtained by dividing the ground coordinate system at a self-defined resolution;

[0098] construct the high-definition map according to the semantic value of each ground grid point.

[0099] The above as described in the present application Figure 1The method performed by the high-definition map construction device disclosed in the embodiment can be applied in a processor or implemented by the processor. The processor can be an integrated circuit chip with signal processing capability. In the implementation process, each step of the above method can be completed by integrated logic circuits of hardware in the processor or instructions in the form of software. The above processor can be a general processor, including a central processing unit (CPU), a network processor (NP), etc.; or a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. Each method, step and logic block disclosed in the embodiment of the present application can be implemented or executed. The general processor can be a microprocessor or any conventional processor. The steps of the method disclosed in combination with the embodiment of the present application can be directly embodied as a hardware decoding processor for execution, or a combination of hardware and software modules in the decoding processor for execution. The software module can be located in a mature storage medium in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and combines the hardware to complete the steps of the above method.

[0100] The electronic device can also execute Figure 1 the method performed by the medium-to-high-definition map construction device, and implement the functions of the high-definition map construction device in Figure 1 the embodiment. The embodiment of the present application will not be repeated here.

[0101] The embodiment of the present application also proposes a computer readable storage medium storing one or more programs, the one or more programs including instructions that, when executed by an electronic device including a plurality of applications, can cause the electronic device to execute Figure 1 the method performed by the medium-to-high-definition map construction device in the embodiment, and specifically for executing

[0102] Obtaining a road element semantic segmentation result from image information collected by a plurality of cameras;

[0103] Correcting the pose transformation relationship of the vehicle relative to the ground coordinate system according to vehicle positioning information, wherein the vehicle positioning information includes positioning information output by a combination navigation device;

[0104] determine a pose transformation relationship of the camera relative to the ground coordinate system according to the pose transformation relationship of the vehicle relative to the ground coordinate system;

[0105] project the ground grid points into the road element semantic segmentation result to obtain a semantic value of each ground grid point according to the pose transformation relationship of the camera relative to the ground coordinate system, wherein the ground grid points are obtained by dividing the ground coordinate system according to a self-defined resolution;

[0106] construct the high-definition map according to the semantic value of each ground grid point.

[0107] Those skilled in the art will understand that embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) containing computer usable program code.

[0108] The present application is described in reference to the flowcharts and / or block diagrams of the methods, apparatus (systems) and computer program products according to embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and combinations of flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, a special purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine, so that the instructions, which are executed via the processor of the computer or other programmable data processing apparatus, generate a means for implementing the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 an apparatus with the specified functions of one or more flows and / or blocks.

[0109] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to work in a specific manner, so that the instructions stored in the computer readable memory produce a manufacture product including instruction apparatus, which implements the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 an apparatus with the specified functions of one or more flows and / or blocks.

[0110] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, so that a series of operation steps are performed on the computer or other programmable data processing apparatus to produce a computer implemented process, so that the instructions executed on the computer or other programmable data processing apparatus provide a means for implementing the functions specified in the flowcharts and / or block diagrams. Figure 1one or more processes and / or blocks ​ steps of a function specified in one or more blocks.

[0111] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0112] The memory can include non-persistent memory and / or volatile memory, such as random access memory (RAM) and / or cache memory, non-volatile memory, such as read-only memory (ROM), EPROM, and / or flash memory, etc. The memory is an example of computer readable media.

[0113] Computer readable media includes permanent and non-permanent, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD), or other optical storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible to a computing device. According to the definition herein, computer readable media does not include transitory media, such as modulated data signals and carrier waves.

[0114] It should also be noted that the terms "comprising," "including," and any other variation thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements recited, but can also include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without limitation, an element preceded by "comprises a" does not, without more constraints, foreclose the existence of additional identical elements in the process, method, article, or apparatus that comprises the recited element.

[0115] Those skilled in the art will appreciate that embodiments of the present application can be devised for a method, a system, or a computer program product. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer readable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) embodying computer readable program code.

[0116] The above embodiments are only used to illustrate the present application, but not to limit it. Various modifications and changes can be made by those skilled in the art. Any modification, equivalent replacement, improvement, and the like made within the spirit and principle of the present application shall fall into the scope of claims of the present application.

Claims

1. A high-precision map construction method, wherein, The method comprises: obtaining a road element semantic segmentation result from image information collected by a camera; correcting a pose transformation relationship of a vehicle relative to a ground coordinate system according to vehicle positioning information, wherein the vehicle positioning information comprises positioning information output by a combined navigation device; determining a pose transformation relationship of the camera relative to the ground coordinate system according to the pose transformation relationship of the vehicle relative to the ground coordinate system; projecting a ground grid point into the road element semantic segmentation result to obtain a semantic value of each ground grid point according to the pose transformation relationship of the camera relative to the ground coordinate system, wherein the ground grid point is obtained by dividing the ground coordinate system at a self-defined resolution; constructing the high-definition map according to the semantic value of each ground grid point.

2. The method of claim 1, wherein, The camera and the combined navigation device are pre-installed on the vehicle, and the method further comprises: simultaneously receiving the image information collected by the camera and the vehicle positioning information output by the combined navigation device, wherein the time stamps of the camera image information and the vehicle positioning information are aligned; performing semantic segmentation of road elements on the image information, and buffering the vehicle positioning information to a preset processing queue, wherein the combined navigation device and the camera-to-vehicle external parameter are pre-calibrated.

3. The method of claim 2, wherein, The correction of the pose transformation relationship of the vehicle relative to the ground coordinate system according to the vehicle positioning information comprises: calculating a relative displacement and a height change corresponding to the vehicle positioning information at the head and tail of the preset processing queue; in the case where the relative displacement is greater than a preset threshold, calculating an angle between the ground and the horizontal plane; calculating a pitch angle of the vehicle relative to the ground according to the pitch angle of the vehicle positioning and the angle between the ground and the horizontal plane; obtaining the pose transformation relationship of the vehicle relative to the ground coordinate system according to the pitch angle of the vehicle relative to the ground and the roll angle of the vehicle positioning.

4. The method of claim 3, wherein, The vehicle positioning information output by the combined navigation device comprises rotation of a vehicle coordinate system relative to a north-eastern sky coordinate system, a planar position in a UTM coordinate system, and an altitude, and the determination of the pose transformation relationship of the camera relative to the ground coordinate system according to the pose transformation relationship of the vehicle relative to the ground coordinate system comprises: obtaining the pose transformation relationship of the camera relative to the ground coordinate system according to the pitch angle of the vehicle relative to the ground and the roll angle of the vehicle positioning, and correcting the roll angle and pitch angle errors of the camera relative to the ground caused by the body fluctuation in real time to determine the pose transformation relationship of the camera relative to the ground coordinate system.

5. The method of claim 1, wherein, The construction of the high-definition map according to the semantic value of each ground grid point comprises: if the semantic value of the ground grid point belongs to a road element, rotating the corresponding ground grid point to a vehicle coordinate system and then updating it to the high-definition map, wherein the road element at least comprises one of the following: lane line, arrow, stop line, and sidewalk.

6. The method of claim 2, wherein, The determination of the pose transformation relationship of the camera relative to the ground coordinate system according to the pose transformation relationship of the vehicle relative to the ground coordinate system comprises: According to the pose transformation relationship of the vehicle relative to the ground coordinate system and the external parameter of the camera to the vehicle body, a pose transformation relationship of the camera relative to the ground coordinate system is determined.

7. The method of claim 1, wherein, The ground grid points are obtained by the following method: In the ground coordinate system, a region of interest is determined according to a self-defined horizontal length and a horizontal width; The region of interest is divided into grids to obtain the ground grid points at a self-defined resolution, and the horizontal height value is 0.

8. A high-definition map construction device in which, The device comprises: A semantic segmentation module configured to obtain a road element semantic segmentation result from image information collected by a plurality of cameras; A correction module configured to correct a pose transformation relationship of a vehicle relative to a ground coordinate system according to vehicle positioning information, wherein the vehicle positioning information comprises positioning information output by a combined navigation device; A determination module configured to determine a pose transformation relationship of a camera relative to the ground coordinate system according to the pose transformation relationship of the vehicle relative to the ground coordinate system; A projection module configured to project the ground grid points into the road element semantic segmentation result to obtain a semantic value of each ground grid point according to the pose transformation relationship of the camera relative to the ground coordinate system, wherein the ground grid points are obtained by dividing the ground coordinate system at a self-defined resolution; A construction module configured to construct the high-definition map according to the semantic value of each ground grid point. 9.An electronic device comprising: a processor; and a memory arranged to store computer executable instructions that, when executed, cause the processor to perform the method of any one of claims 1-7. 10.A computer readable storage medium storing one or more programs, the one or more programs, when executed by an electronic device including multiple applications, cause the electronic device to perform the method of any one of claims 1-7. ​

Images