A high-precision map processing method and device

Abstract

Embodiments of the present application relate to a kind of high-precision map processing method and device, the method comprises: in the process of first vehicle driving, the road ahead is surveyed and is surveyed and is generated first survey view;First survey view is detected to generate multiple first target detection frame;According to all first target detection frame, confirm whether the road scene in front is road intersection scene;If it is confirmed as road intersection scene, then sort out map information calibration board detection frame and positioning calibration board detection frame;According to map information calibration board detection frame, standard high-precision map acquisition is obtained first standard high-precision map;According to all positioning calibration board detection frame and first standard high-precision map, self car map coordinate confirmation is generated first vehicle map coordinate;According to first vehicle map coordinate and all first target detection frame, first standard high-precision map is fused to generate first intersection high-precision map.The present application is helpful to improve the response speed of vehicle automatic driving system in road intersection.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method and apparatus for processing high-precision maps. Background Technology

[0002] Autonomous vehicles require timely updates to high-precision maps of their location, especially at complex intersections. Currently, the conventional method for updating these maps involves obtaining the vehicle's latitude and longitude parameters through its positioning module and then querying a high-precision map database based on these parameters. This approach has several drawbacks: 1) When the positioning module fails to locate the vehicle or has significant errors, the high-precision map may not be updated promptly or may be inaccurate; 2) When the vehicle is at an intersection, the high-precision map obtained is merely a standard map without real-time obstacle information. These issues all affect the response speed of the autonomous driving system at intersections. Summary of the Invention

[0003] The purpose of this invention is to address the shortcomings of existing technologies by providing a high-precision map processing method, apparatus, electronic device, and computer-readable storage medium. The method involves pre-installing a map information calibration board and at least two positioning calibration boards on the side of each road leading into the intersection. During vehicle travel, real-time information of all obstacle targets is obtained through target detection in a surrounding view of the road ahead. The type of each obstacle target is identified to confirm whether an intersection is approaching. Upon confirmation of approaching the intersection, QR code information on the map information calibration board and positioning calibration board in the surrounding view is identified to obtain corresponding standard high-precision map and positioning calibration board identifiers. The position of the corresponding positioning calibration board is confirmed on the standard high-precision map based on the positioning calibration board identifiers. Real-time vehicle semantic information is integrated into the standard high-precision map based on the relative displacement relationship between the positioning calibration board position and the vehicle, and real-time obstacle target semantic information is integrated into the standard high-precision map based on the relative displacement relationship between the vehicle and each obstacle target. This results in a real-world high-precision map of the intersection that integrates real-time information of the vehicle and obstacle targets. This invention can solve the problem of untimely acquisition of high-precision maps caused by positioning module failure, and can provide a high-precision intersection map that integrates real-world traffic conditions, which helps improve the response speed of vehicle autonomous driving systems at road intersections.

[0004] To achieve the above objectives, a first aspect of the present invention provides a method for processing high-precision maps, the method comprising:

[0005] During the first vehicle's journey, a surround view image is taken of the road ahead to generate the corresponding first surround view.

[0006] The first ring view is subjected to target detection processing to generate multiple first target detection boxes;

[0007] Based on all the first target detection boxes, it is confirmed whether the road scene ahead is a road intersection scene; if it is confirmed that the road scene ahead is a road intersection scene, the map information calibration board detection box and the positioning calibration board detection box are selected from the multiple first target detection boxes.

[0008] Based on the map information calibration board detection box, a standard high-precision map acquisition process is performed to generate the corresponding first standard high-precision map.

[0009] Based on all the positioning calibration board detection frames and the first standard high-precision map, the vehicle map coordinates are confirmed and the corresponding first vehicle map coordinates are generated.

[0010] Based on the map coordinates of the first vehicle and all the first target detection boxes, the first standard high-precision map is subjected to intersection target fusion processing to generate the corresponding first intersection high-precision map.

[0011] Preferably, before taking a surround view of the road ahead during the first vehicle's movement to generate the corresponding first surround view, the method further includes:

[0012] One map information calibration board and at least two positioning calibration boards are pre-installed on the side of each road leading into the intersection; the map information calibration board and the positioning calibration board are rectangular by default; a QR code containing calibration board type information, map information calibration board identification information and standard high-precision map information is pre-drawn on the map information calibration board; a QR code containing the calibration board type information and positioning calibration board identification information is pre-drawn on the positioning calibration board; the calibration board type information includes map information calibration board type and positioning calibration board type.

[0013] Preferably, the step of performing target detection processing on the first ring view to generate multiple first target detection boxes specifically includes:

[0014] Based on a preset image target detection model, image target detection processing is performed on the first surround view to obtain multiple first target detection boxes; the first target detection box includes the coordinates of the center point of the first detection box, the orientation angle of the first detection box, the size of the first detection box, and the target type of the first detection box; the target types of the first detection box include vehicle type, animal type, human type, traffic obstacle type, building type, calibration board type, and traffic light type; the coordinate system of the center point coordinates of the first detection box is the vehicle coordinate system.

[0015] Preferably, the step of confirming whether the road scene ahead is a road intersection scene based on all the first target detection boxes specifically includes:

[0016] Check whether there is a traffic light type among the target types of all the first target detection boxes obtained; if it is confirmed that there is, then the road scene ahead is confirmed as a road intersection scene; if it is confirmed that there is no, then the road scene ahead is confirmed as a non-road intersection scene.

[0017] Preferably, the step of selecting the map information calibration board detection box and the positioning calibration board detection box from the obtained plurality of first target detection boxes specifically includes:

[0018] The first target detection box whose target type is calibration board type is denoted as the corresponding first calibration board detection box;

[0019] Extract the sub-images on the first annular view that correspond to each of the first calibration plate detection boxes as the corresponding first calibration plate images.

[0020] Each first calibration board image is processed by QR code information recognition to obtain corresponding first QR code information; the first QR code information includes calibration board type information; the calibration board type information includes map information calibration board type and positioning calibration board type; when the calibration board type information is map information calibration board type, the first QR code information is composed of the calibration board type information, map information calibration board identification information, and standard high-precision map information; when the calibration board type information is positioning calibration board type, the first QR code information is composed of the calibration board type information and positioning calibration board identification information;

[0021] The first target detection box corresponding to the first calibration board image whose calibration board type information is map information calibration board type is used as the corresponding map information calibration board detection box; and the first target detection box corresponding to the first calibration board image whose calibration board type information is positioning calibration board type is used as the corresponding positioning calibration board detection box.

[0022] Preferably, the step of generating a corresponding first standard high-precision map by performing standard high-precision map acquisition processing based on the map information calibration board detection box specifically includes:

[0023] The standard high-precision map information of the first QR code information corresponding to the detection box of the map information calibration board is taken as the corresponding first standard high-precision map; the map coordinate system of the first standard high-precision map is the road intersection coordinate system; the first standard high-precision map is a semantic map, and the semantic information of the first standard high-precision map includes road semantic information, road edge semantic information, lane semantic information, lane center line semantic information, lane edge line semantic information, traffic light semantic information, positioning signboard semantic information, and pedestrian semantic information; the positioning signboard semantic information includes positioning signboard semantic type, positioning signboard identifier, and positioning signboard coordinates.

[0024] Preferably, the step of generating the corresponding first vehicle map coordinates by confirming the vehicle map coordinates based on all the positioning calibration board detection frames and the first standard high-precision map specifically includes:

[0025] The straight-line distance between the current positioning calibration plate detection frame and the workshop is calculated based on the coordinates of the center point of the first detection frame of each of the positioning calibration plate detection frames, and is taken as the corresponding first distance d.

[0026] The positioning calibration board identifier information of the first QR code information corresponding to each positioning calibration board detection box is used as the corresponding second identifier; and the positioning signboard coordinates of the positioning signboard semantic information corresponding to the positioning signboard identifier that matches each of the second identifiers on the first standard high-precision map are extracted as the corresponding first positioning coordinates.

[0027] On the first standard high-precision map, a circle is drawn with the first positioning coordinates of each of the positioning calibration board detection boxes as the center and the corresponding first spacing d as the radius to obtain the corresponding first circle; the common intersection point of all the first circles is taken as the corresponding first intersection point; and the number of the first intersection points is counted to generate the corresponding number of first intersection points; the number of the first intersection points is greater than or equal to 1.

[0028] When the number of the first intersection points is 1, the map coordinates of the first intersection point on the first standard high-precision map are used as the corresponding map coordinates of the first vehicle and output.

[0029] When the number of first intersection points is greater than 1, each of the first intersection points is traversed. During traversal, the currently traversed first intersection point is taken as the corresponding current intersection point. A corresponding vehicle coordinate system is constructed with the current intersection point as the origin of the vehicle coordinate system and recorded as the corresponding current vehicle coordinate system. The first positioning coordinates of each positioning calibration board detection frame are transformed from the road intersection coordinate system to the current vehicle coordinate system to obtain the corresponding second positioning coordinates. The coordinate deviation between the center point coordinates of the first detection frame and the second positioning coordinates of each positioning calibration board detection frame is calculated to generate the corresponding first coordinate deviation. It is then identified whether the first coordinate deviations corresponding to all the obtained positioning calibration board detection frames are lower than the preset minimum coordinate deviation threshold. If not, the traversal continues to the next first intersection point until the traversal of the last first intersection point ends. If yes, the traversal stops and the map coordinates of the current intersection point on the first standard high-precision map are output as the corresponding first vehicle map coordinates.

[0030] Preferably, the step of performing intersection target fusion processing on the first standard high-precision map based on the first vehicle map coordinates and all the first target detection boxes to generate a corresponding first intersection high-precision map specifically includes:

[0031] A corresponding vehicle coordinate system is constructed using the first vehicle map coordinates as the origin of the vehicle coordinate system, and denoted as the current vehicle coordinate system. The coordinate transformation relationship between the current vehicle coordinate system and the road intersection coordinate system of the first standard high-precision map is confirmed to obtain the corresponding first transformation relationship. Based on the first transformation relationship, the coordinates of the center point of the first detection box and the orientation angle of the first detection box of each first target detection box are corrected to generate the corresponding second detection box center point coordinates and second detection box orientation angle.

[0032] On the first standard high-precision map, semantic information of the vehicle is added based on the vehicle orientation, shape and size of the first vehicle and the map coordinates of the first vehicle; and semantic information of the obstacle target is added based on the coordinates of the center point of the second detection box, the orientation angle of the second detection box, the size of the first detection box and the target type of the first detection box of each first target detection box.

[0033] The first standard high-precision map, after completing the semantic information addition processing for the vehicle and all obstacle targets, is output as the corresponding high-precision map of the first intersection.

[0034] A second aspect of the present invention provides an apparatus for implementing the high-precision map processing method described in the first aspect above, the apparatus comprising: a panoramic view processing module, a target detection processing module, a standard high-precision map processing module, and a high-precision map fusion processing module;

[0035] The surround view processing module is used to take a surround view picture of the road ahead during the first vehicle's driving process to generate a corresponding first surround view.

[0036] The target detection processing module is used to perform target detection processing on the first ring view to generate multiple first target detection boxes;

[0037] The standard high-precision map processing module is used to confirm whether the road scene ahead is a road intersection scene based on all the first target detection boxes; if the road scene ahead is confirmed to be a road intersection scene, the map information calibration board detection box and the positioning calibration board detection box are selected from the multiple first target detection boxes; and the standard high-precision map acquisition processing is performed based on the map information calibration board detection box to generate the corresponding first standard high-precision map;

[0038] The high-precision map fusion processing module is used to perform vehicle map coordinate confirmation processing based on all the positioning calibration board detection boxes and the first standard high-precision map to generate the corresponding first vehicle map coordinates; and to perform intersection target fusion processing on the first standard high-precision map based on the first vehicle map coordinates and all the first target detection boxes to generate the corresponding first intersection high-precision map.

[0039] A third aspect of the present invention provides an electronic device, including: a memory, a processor, and a transceiver;

[0040] The processor is used to couple with the memory, read and execute instructions in the memory to implement the steps of the method described in the first aspect above;

[0041] The transceiver is coupled to the processor, and the processor controls the transceiver to send and receive messages.

[0042] A fourth aspect of the present invention provides a computer-readable storage medium storing computer instructions that, when executed by a computer, cause the computer to perform the instructions described in the first aspect.

[0043] This invention provides a method, apparatus, electronic device, and computer-readable storage medium for processing high-precision maps. The method involves pre-installing a map information calibration board and at least two positioning calibration boards on the side of each road leading into the intersection. During vehicle travel, real-time information of all obstacle targets is obtained by performing target detection on a surrounding view of the road ahead. The type of each obstacle target is identified to confirm whether an intersection is approaching. Upon confirmation of approaching the intersection, QR code information on the map information calibration board and positioning calibration board in the surrounding view is identified to obtain corresponding standard high-precision map and positioning calibration board identifiers. The position of the corresponding positioning calibration board is confirmed on the standard high-precision map based on the positioning calibration board identifier. Real-time vehicle semantic information is integrated into the standard high-precision map based on the relative displacement relationship between the positioning calibration board position and the vehicle, and real-time obstacle target semantic information is integrated into the standard high-precision map based on the relative displacement relationship between the vehicle and each obstacle target. This results in a real-scene high-precision map of the intersection that integrates real-time information of the vehicle and obstacle targets. This invention solves the problem of untimely acquisition of high-precision maps caused by positioning module failure, and provides a high-precision intersection map that integrates real-world traffic conditions, thereby improving the response speed of the vehicle's autonomous driving system at road intersections. Attached Figure Description

[0044] Figure 1 This is a schematic diagram of a high-precision map processing method provided in Embodiment 1 of the present invention;

[0045] Figure 2 This is a module structure diagram of a high-precision map processing device provided in Embodiment 2 of the present invention;

[0046] Figure 3 This is a schematic diagram of the structure of an electronic device provided in Embodiment 3 of the present invention. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0048] Embodiment 1 of the present invention provides a method for processing high-precision maps, such as... Figure 1 The diagram illustrates a high-precision map processing method provided in Embodiment 1 of the present invention. This method mainly includes the following steps:

[0049] Step 1: During the first vehicle's journey, take a panoramic view of the road ahead to generate the corresponding first panoramic view;

[0050] Here, the first vehicle in Embodiment 1 of the present invention is a motor vehicle that is undergoing autonomous or driverless operation, and the term "autonomous vehicle" in the following text also refers to the first vehicle; the first vehicle has an autonomous driving system for planning autonomous or driverless operation; the autonomous driving system of Embodiment 1 of the present invention will call the camera installed on the first vehicle to continuously take panoramic view pictures of the road in front of the vehicle at a preset shooting frequency during the first vehicle's operation, and take the panoramic view obtained each time as the first panoramic view obtained at that time; the first panoramic view here can be a panoramic view stitched together from multiple angle images taken simultaneously by multiple cameras with different shooting angles by the autonomous driving system, or it can be a panoramic view stitched together from multiple angle images taken by a single camera by rotating it to different shooting angles by the autonomous driving system.

[0051] It should be noted that, before step 1 above, which involves taking a surround view of the road ahead during driving to generate the corresponding first surround view, the method of Embodiment 1 of the present invention further includes:

[0052] Pre-install one map information calibration board and at least two positioning calibration boards on the side of each road leading into the intersection. The map information calibration board and positioning calibration board are rectangular by default. Each map information calibration board has a pre-drawn QR code containing calibration board type information, map information calibration board identification information, and standard high-precision map information. Each positioning calibration board also has a pre-drawn QR code containing calibration board type information and positioning calibration board identification information. The calibration board type information includes both map information calibration board type and positioning calibration board type. These two types of calibration boards will be used in subsequent steps for standard high-precision map acquisition and vehicle coordinate positioning. It should be noted that if two positioning calibration boards are pre-installed, they are installed on both sides of the corresponding road by default. If more than two positioning calibration boards are pre-installed, at least one positioning calibration board should be installed on both sides of the corresponding road.

[0053] Step 2: Perform target detection processing on the first ring view to generate multiple first target detection boxes;

[0054] Specifically, this includes: performing image target detection processing on the first ring view based on a preset image target detection model to obtain multiple first target detection boxes;

[0055] The first target detection box includes the coordinates of the center point of the first detection box, the orientation angle of the first detection box, the size of the first detection box, and the target type of the first detection box; the target type of the first detection box includes vehicle type, animal type, human type, traffic obstacle type, building type, calibration board type, and traffic light type; the coordinate system of the center point coordinates of the first detection box is the vehicle coordinate system.

[0056] Here, the image target detection model used in the autonomous driving system of Embodiment 1 of the present invention is a pre-trained and mature model that can be used to detect vehicles, animals, people, traffic obstacles, buildings, calibration boards, and traffic lights. The output structure of the image target detection model consists of one or more first target detection boxes (bboxes), each of which corresponds to a detected obstacle target. The center point coordinates, orientation angle, and size of the first detection box are the center point coordinates, orientation angle, and size (such as height and width) of the corresponding target detection box, respectively. The target type of the first detection box is the obstacle target type of the corresponding target detection box.

[0057] The image target detection model of Embodiment 1 of this invention has multiple implementation methods, including: an image target detection model based on the Faster RCNN model structure, an image target detection model based on the YO10 series model structure, and an image target detection model based on the SSD model structure. When training the image target detection model, a large number of panoramic views containing obstacle targets such as vehicles, animals, people, traffic obstacles, buildings, calibration boards, or traffic lights are collected in advance as training data to train the image target detection model until the detection error between each target detection box output by the model and each obstacle target such as vehicle, animal, person, traffic obstacle, building, calibration board, or traffic light on the panoramic view converges to a specified error range.

[0058] Step 3: Confirm whether the road scene ahead is a road intersection scene based on all the first target detection boxes; if it is confirmed that the road scene ahead is a road intersection scene, then select the map information calibration board detection box and the positioning calibration board detection box from the multiple first target detection boxes obtained.

[0059] Specifically, this includes: Step 31, confirming whether the road scene ahead is a road intersection scene based on all the first target detection boxes;

[0060] Specifically, this includes: confirming whether a traffic light type exists among the target types of all the first target detection boxes obtained; if it is confirmed to exist, then confirming that the road scene ahead is a road intersection scene; if it is confirmed not to exist, then confirming that the road scene ahead is not a road intersection scene.

[0061] Here, if it is confirmed that there is a traffic light type among the target types of all first target detection boxes, it means that the probability of a traffic intersection is appearing on the road ahead is very high. At this time, the autonomous driving system of Embodiment 1 of the present invention confirms that the road scene ahead is a road intersection scene and continues to execute the subsequent steps; conversely, if it is confirmed that there is no traffic light type among the target types of all first target detection boxes, it means that the probability of a traffic intersection is appearing on the road ahead is very low. At this time, the autonomous driving system of Embodiment 1 of the present invention confirms that the road scene ahead is not a road intersection scene and stops executing the subsequent steps.

[0062] Step 32: If it is confirmed that the road scene ahead is a road intersection scene, then the map information calibration board detection box and the positioning calibration board detection box are selected from the multiple first target detection boxes.

[0063] Specifically, this includes: step 321, recording the first target detection box whose target type is calibration board type as the corresponding first calibration board detection box;

[0064] Step 322: Extract the sub-images corresponding to each detection box of the first calibration plate on the first ring view as the corresponding first calibration plate images;

[0065] Step 323: Perform QR code information recognition processing on each first calibration board image to obtain the corresponding first QR code information;

[0066] The first QR code information includes calibration board type information; the calibration board type information includes map information calibration board type and positioning calibration board type; when the calibration board type information is map information calibration board type, the first QR code information consists of calibration board type information, map information calibration board identification information and standard high-precision map information; when the calibration board type information is positioning calibration board type, the first QR code information consists of calibration board type information and positioning calibration board identification information.

[0067] Here, in the first embodiment of the present invention, when the autonomous driving system performs QR code information recognition processing on each first calibration board image, it calls a preset QR code image processing interface to perform QR code image filtering, calibration, and QR code encoding recognition on each first calibration board image to obtain the corresponding QR code recognition information, i.e., the first QR code information. The implementation process of the QR code image processing interface here adopts conventional QR code image recognition technology, which will not be described in detail here. The QR code system here can also be specified according to specific implementation requirements, which will not be described in detail here either.

[0068] It should be noted that, as shown above, a QR code containing calibration board type information, map information calibration board identification information, and standard high-precision map information is pre-drawn on the map information calibration board, and a QR code containing calibration board type information and positioning calibration board identification information is pre-drawn on the positioning calibration board. Therefore, the first QR code information identified in the current step consists of calibration board type information, map information calibration board identification information, and standard high-precision map information when the calibration board type information is map information calibration board type, and consists of calibration board type information and positioning calibration board identification information when the calibration board type information is positioning calibration board type.

[0069] Step 324: The first target detection box corresponding to the first calibration board image with calibration board type information of map information calibration board type is used as the corresponding map information calibration board detection box; and the first target detection box corresponding to the first calibration board image with calibration board type information of positioning calibration board type is used as the corresponding positioning calibration board detection box.

[0070] Here, as mentioned earlier, a map information calibration board and at least two positioning calibration boards will be installed on the side of each road at each intersection. Therefore, the number of detection boxes for the map information calibration board should be 1, while the number of detection boxes for the positioning calibration board should be greater than or equal to 2.

[0071] Step 4: Based on the map information calibration board detection box, perform standard high-precision map acquisition processing to generate the corresponding first standard high-precision map;

[0072] Specifically, this includes: using the standard high-precision map information of the first QR code information corresponding to the map information calibration board detection box as the corresponding first standard high-precision map;

[0073] Among them, the map coordinate system of the first standard high-precision map is the road intersection coordinate system; the first standard high-precision map is a semantic map, and the semantic information of the first standard high-precision map includes road semantic information, road edge semantic information, lane semantic information, lane centerline semantic information, lane edgeline semantic information, traffic light semantic information, positioning signboard semantic information, and pedestrian semantic information; the positioning signboard semantic information includes positioning signboard semantic type, positioning signboard identifier, and positioning signboard coordinates.

[0074] It should be noted that the standard high-precision map information in Embodiment 1 of the present invention is actually the semantic map vector data of a standard high-precision map. The map vector encoding rules of the semantic map vector data are known. These rules can be conventional map vector encoding rules or customized according to application requirements, which will not be elaborated further here.

[0075] Step 5: Based on all the positioning calibration board detection frames and the first standard high-precision map, perform self-vehicle map coordinate confirmation processing to generate the corresponding first vehicle map coordinates;

[0076] Specifically, it includes: Step 51, calculating the straight-line distance between the current positioning calibration plate detection frame and the workshop based on the coordinates of the center point of the first detection frame of each positioning calibration plate detection frame as the corresponding first distance d;

[0077] Here, since the origin is always the vehicle in the vehicle coordinate system, the first spacing d is actually the straight-line distance from the center point of the first detection box to the origin in the vehicle coordinate system.

[0078] For example, if the coordinates of the center point of the first detection box are (x, y) in a two-dimensional vehicle coordinate system, then... For example, if the coordinates of the center point of the first detection frame are (x, y, z) in the three-dimensional vehicle coordinate system, then...

[0079] Step 52: Take the positioning calibration board identifier information of the first QR code information corresponding to each positioning calibration board detection box as the corresponding second identifier; and extract the positioning signboard coordinates of the positioning signboard semantic information corresponding to the positioning signboard identifier that matches each second identifier on the first standard high-precision map as the corresponding first positioning coordinates.

[0080] Here, each first positioning coordinate is actually the positioning coordinate of each positioning signboard in the road intersection coordinate system of the first standard high-precision map;

[0081] Step 53: On the first standard high-precision map, draw a circle with the first positioning coordinates of each positioning calibration plate detection box as the center and the corresponding first spacing d as the radius to obtain the corresponding first circle; take the common intersection point of all the first circles as the corresponding first intersection point; and count the number of the first intersection points to generate the corresponding number of first intersection points.

[0082] Among them, the number of first intersection points is greater than or equal to 1;

[0083] Here, the most common value of the number of first intersection points in practical applications is either 1 or 2, and it is rare for it to be greater than 2.

[0084] Step 54: When the number of first intersection points is 1, the map coordinates of the first intersection point on the first standard high-precision map are used as the corresponding map coordinates of the first vehicle and output.

[0085] Here, if the number of first intersection points is equal to 1, it means that this unique first intersection point is the actual coordinate point of the vehicle on the first standard high-precision map. Therefore, the map coordinates of this unique first intersection point on the first standard high-precision map are output as the corresponding first vehicle map coordinates.

[0086] Step 55: When the number of first intersection points is greater than 1, traverse each first intersection point; during traversal, the currently traversed first intersection point is taken as the corresponding current intersection point; and a corresponding vehicle coordinate system is constructed with the current intersection point as the origin of the vehicle coordinate system, denoted as the corresponding current vehicle coordinate system; and the first positioning coordinates of each positioning calibration board detection frame are transformed from the road intersection coordinate system to the current vehicle coordinate system to obtain the corresponding second positioning coordinates; and the coordinate deviation between the center point coordinates of the first detection frame and the second positioning coordinates of each positioning calibration board detection frame is calculated to generate the corresponding first coordinate deviation; and it is identified whether the first coordinate deviations corresponding to all the obtained positioning calibration board detection frames are lower than the preset minimum coordinate deviation threshold; if not, proceed to the next first intersection point and continue traversing until the traversal of the last first intersection point ends; if yes, stop traversing and output the map coordinates of the current intersection point on the first standard high-precision map as the corresponding first vehicle map coordinates.

[0087] Here, if the number of first intersection points is greater than 1, it is necessary to verify one by one whether each first intersection point is the actual coordinate point of the vehicle on the first standard high-precision map. Therefore, it is necessary to traverse each first intersection point. During the traversal, the position of the currently traversed first intersection point is taken as the current vehicle position, and a vehicle coordinate system is constructed based on the current vehicle position. The positioning coordinates of each positioning calibration board, i.e., each positioning calibration board detection frame, on the first standard high-precision map (i.e., the first positioning coordinates), are transformed to the current vehicle coordinate system to obtain the corresponding current vehicle coordinates, i.e., the second positioning coordinates. In principle, if the currently traversed first intersection point is the actual coordinate point of the vehicle on the first standard high-precision map, then the second positioning coordinates of each positioning calibration board should correspond to the coordinates of the center point of the corresponding first detection frame. While they overlap, conversion errors may occur during actual calculations. Therefore, in Embodiment 1 of this invention, a small error threshold, i.e., the minimum coordinate deviation threshold, is preset. As long as the coordinate deviation between the second positioning coordinates corresponding to each positioning calibration board and the coordinates of the center point of the first detection frame, i.e., the first coordinate deviation, is less than the minimum coordinate deviation threshold, they are considered to overlap. Therefore, after obtaining the first coordinate deviations of all positioning calibration boards, it is necessary to identify whether all the obtained first coordinate deviations are lower than the preset minimum coordinate deviation threshold. If so, it means that the first intersection point of the current traversal is the real coordinate point of the vehicle on the first standard high-precision map. At this time, the traversal is immediately ended, and the map coordinates of the first intersection point of the current traversal, i.e., the current intersection point, on the first standard high-precision map are output as the corresponding first vehicle map coordinates.

[0088] Step 6: Perform intersection target fusion processing on the first standard high-precision map based on the first vehicle map coordinates and all first target detection boxes to generate the corresponding first intersection high-precision map;

[0089] Specifically, this includes: Step 61, constructing a corresponding vehicle coordinate system with the first vehicle map coordinates as the origin of the vehicle coordinate system, denoted as the current vehicle coordinate system; confirming the coordinate transformation relationship between the current vehicle coordinate system and the road intersection coordinate system of the first standard high-precision map to obtain the corresponding first transformation relationship; and correcting the coordinates of the center point of the first detection box and the orientation angle of the first detection box of each first target detection box according to the first transformation relationship to generate the corresponding coordinates of the center point of the second detection box and the orientation angle of the second detection box;

[0090] Here, the current vehicle coordinate system is a two-dimensional or three-dimensional coordinate system with the vehicle as the origin, while the road intersection coordinate system of the first standard high-precision map is a two-dimensional or three-dimensional coordinate system with the center point of the road intersection as the origin. The transformation relationship between the two, i.e., the first transformation relationship, can be regarded as a transformation relationship vector composed of two or three coordinate components of the vehicle's coordinates in the road intersection coordinate system, i.e., the first vehicle map coordinates. Based on this transformation relationship vector, transforming the coordinates of the detection box center point in the vehicle coordinate system, i.e., the coordinates of the first detection box center point, yields the corresponding coordinates in the road intersection coordinate system, i.e., the coordinates of the second detection box center point. Based on this transformation relationship vector, adjusting the angle of the detection box orientation in the vehicle coordinate system yields the corresponding orientation angle in the road intersection coordinate system, i.e., the orientation angle of the second detection box. Because this coordinate and orientation angle transformation between the two coordinate systems is a publicly available and conventional transformation technique, the specific calculation process will not be elaborated further here.

[0091] Step 62: On the first standard high-precision map, add semantic information of the vehicle based on the vehicle orientation, shape and size, and map coordinates of the first vehicle; and add semantic information of obstacle targets based on the center point coordinates of the second detection box, the orientation angle of the second detection box, the size of the first detection box, and the target type of the first detection box for each first target detection box.

[0092] Here, after obtaining the coordinates, orientation, shape, and size of the vehicle and each obstacle target on the first standard high-precision map, the map space matching the vehicle and each obstacle target can be determined on the first standard high-precision map. Adding the vehicle's (i.e., the first vehicle's) vehicle orientation, shape, size, and map coordinates to the semantic information set of the map space matching the vehicle completes the addition of the vehicle's semantic information to the first standard high-precision map. Adding the coordinates of the center point of the second detection box, the orientation angle of the second detection box, the size of the first detection box, and the target type of the first detection box to the semantic information set of the map space matching each obstacle target completes the addition of the obstacle target semantic information to the first standard high-precision map.

[0093] Step 63: Output the first standard high-precision map, after completing the semantic information addition processing for the vehicle and all obstacle targets, as the corresponding high-precision map of the first intersection.

[0094] Here, the final high-precision map of the first intersection is a high-precision map of the intersection that integrates the real-world traffic conditions. The autonomous driving system of Embodiment 1 of this invention can then use this high-precision map of the first intersection to make vehicle driving strategy decisions and plan driving trajectories, thereby shortening decision-making time and accelerating trajectory planning, thus achieving the goal of improving system response speed at road intersections.

[0095] Figure 2 This is a module structure diagram of a high-precision map processing device provided in Embodiment 2 of the present invention. This device can be a terminal device or server implementing the aforementioned method embodiments, or it can be a device that enables the aforementioned terminal device or server to implement the aforementioned method embodiments. For example, the device can be a device or chip system of the aforementioned terminal device or server. Figure 2 As shown, the device includes: a panoramic view processing module 201, a target detection processing module 202, a standard high-precision map processing module 203, and a high-precision map fusion processing module 204.

[0096] The surround view processing module 201 is used to take surround view pictures of the road ahead during the first vehicle's travel to generate the corresponding first surround view.

[0097] The target detection processing module 202 is used to perform target detection processing on the first ring view to generate multiple first target detection boxes.

[0098] The standard high-precision map processing module 203 is used to confirm whether the road scene ahead is a road intersection scene based on all the first target detection boxes; if it is confirmed that the road scene ahead is a road intersection scene, the map information calibration board detection box and the positioning calibration board detection box are selected from the multiple first target detection boxes; and the standard high-precision map acquisition processing is performed based on the map information calibration board detection box to generate the corresponding first standard high-precision map.

[0099] The high-precision map fusion processing module 204 is used to perform vehicle map coordinate confirmation processing based on all positioning calibration board detection boxes and the first standard high-precision map to generate the corresponding first vehicle map coordinates; and to perform intersection target fusion processing on the first standard high-precision map based on the first vehicle map coordinates and all first target detection boxes to generate the corresponding first intersection high-precision map.

[0100] The high-precision map processing device provided in this embodiment of the invention can execute the method steps in the above method embodiment. Its implementation principle and technical effect are similar, and will not be repeated here.

[0101] It should be noted that the division of the various modules in the above device is merely a logical functional division. In actual implementation, they can be fully or partially integrated into a single physical entity, or they can be physically separated. Furthermore, these modules can be implemented entirely in software via processing element calls; they can be fully implemented in hardware; or some modules can be implemented by processing element calls to software, while others are implemented in hardware. For example, the target detection processing module can be a separate processing element, or it can be integrated into a chip in the above device. Alternatively, it can be stored as program code in the memory of the above device, and called and executed by a processing element of the device. The implementation of other modules is similar. Moreover, these modules can be fully or partially integrated together, or they can be implemented independently. The processing element described here can be an integrated circuit with signal processing capabilities. In the implementation process, each step of the above method or each of the above modules can be completed through integrated logic circuits in the hardware of the processor element or through software instructions.

[0102] For example, these modules can be one or more integrated circuits configured to implement the above methods, such as one or more Application Specific Integrated Circuits (ASICs), one or more Digital Signal Processors (DSPs), or one or more Field Programmable Gate Arrays (FPGAs). Furthermore, when a module is implemented using processing element scheduler code, the processing element can be a general-purpose processor, such as a Central Processing Unit (CPU) or other processor capable of calling program code. Additionally, these modules can be integrated together as a System-on-a-Chip (SOC).

[0103] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. This computer program product includes one or more computer instructions. When these computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the foregoing method embodiments are generated. The computer described above can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The aforementioned computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the aforementioned computer instructions can be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL)) or wireless (e.g., infrared, wireless, Bluetooth, microwave, etc.) means. The aforementioned computer-readable storage medium can be any available medium that a computer can access, or a data storage device such as a server or data center that integrates one or more available media. The aforementioned available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid-state drives (SSDs)).

[0104] Figure 3 This is a schematic diagram of an electronic device provided in Embodiment 3 of the present invention. This electronic device can be the aforementioned terminal device or server, or it can be a terminal device or server connected to the aforementioned terminal device or server that implements the method of the embodiments of the present invention. Figure 3 As shown, the electronic device may include: a processor 301 (e.g., CPU), a memory 302, and a transceiver 303; the transceiver 303 is coupled to the processor 301, and the processor 301 controls the transmission and reception operations of the transceiver 303. The memory 302 may store various instructions for performing various processing functions and implementing the processing steps described in the foregoing method embodiments. Preferably, the electronic device involved in the embodiments of the present invention further includes: a power supply 304, a system bus 305, and a communication port 306. The system bus 305 is used to realize communication connections between components. The communication port 306 is used for communication between the electronic device and other peripherals.

[0105] exist Figure 3The system bus 305 mentioned can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This system bus can be divided into address bus, data bus, control bus, etc. For ease of representation, it is represented by only one thick line in the figure, but this does not indicate that there is only one bus or one type of bus. The communication interface is used to enable communication between the database access device and other devices (e.g., clients, read-write libraries, and read-only libraries). Memory may include Random Access Memory (RAM) and may also include Non-Volatile Memory, such as at least one disk storage device.

[0106] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), graphics processing units (GPUs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0107] It should be noted that the embodiments of the present invention also provide a computer-readable storage medium storing instructions that, when run on a computer, cause the computer to perform the methods and processes provided in the above embodiments.

[0108] This invention also provides a chip for executing instructions, which is used to perform the processing steps described in the foregoing method embodiments.

[0109] This invention provides a method, apparatus, electronic device, and computer-readable storage medium for processing high-precision maps. The method involves pre-installing a map information calibration board and at least two positioning calibration boards on the side of each road leading into the intersection. During vehicle travel, real-time information of all obstacle targets is obtained by performing target detection on a surrounding view of the road ahead. The type of each obstacle target is identified to confirm whether an intersection is approaching. Upon confirmation of approaching the intersection, QR code information on the map information calibration board and positioning calibration board in the surrounding view is identified to obtain corresponding standard high-precision map and positioning calibration board identifiers. The position of the corresponding positioning calibration board is confirmed on the standard high-precision map based on the positioning calibration board identifier. Real-time vehicle semantic information is integrated into the standard high-precision map based on the relative displacement relationship between the positioning calibration board position and the vehicle, and real-time obstacle target semantic information is integrated into the standard high-precision map based on the relative displacement relationship between the vehicle and each obstacle target. This results in a real-scene high-precision map of the intersection that integrates real-time information of the vehicle and obstacle targets. This invention solves the problem of untimely acquisition of high-precision maps caused by positioning module failure, and provides a high-precision intersection map that integrates real-world traffic conditions, thereby improving the response speed of the vehicle's autonomous driving system at road intersections.

[0110] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0111] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented in hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0112] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for processing high-precision maps, characterized in that, The method includes: During the first vehicle's journey, a surround view image is taken of the road ahead to generate the corresponding first surround view. The first ring view is subjected to target detection processing to generate multiple first target detection boxes; Based on all the first target detection boxes, it is confirmed whether the road scene ahead is a road intersection scene; if it is confirmed that the road scene ahead is a road intersection scene, the map information calibration board detection box and the positioning calibration board detection box are selected from the multiple first target detection boxes. Based on the map information calibration board detection box, standard high-precision map acquisition processing is performed to generate the corresponding first standard high-precision map; Based on all the positioning calibration board detection frames and the first standard high-precision map, the vehicle map coordinates are confirmed and the corresponding first vehicle map coordinates are generated. Based on the first vehicle map coordinates and all the first target detection boxes, the first standard high-precision map is subjected to intersection target fusion processing to generate the corresponding first intersection high-precision map. Specifically, the step of performing target detection processing on the first ring view to generate multiple first target detection boxes includes: Based on a preset image target detection model, image target detection processing is performed on the first surround view to obtain multiple first target detection boxes; each first target detection box includes the coordinates of the center point of the first detection box, the orientation angle of the first detection box, the size of the first detection box, and the target type of the first detection box; the target types of the first detection boxes include vehicle type, animal type, human type, traffic obstacle type, building type, calibration board type, and traffic light type; the coordinate system of the center point coordinates of the first detection box is the vehicle coordinate system; The step of selecting map information calibration board detection boxes and positioning calibration board detection boxes from the obtained plurality of first target detection boxes specifically includes: The first target detection box whose target type is calibration board type is denoted as the corresponding first calibration board detection box; Extract the sub-images on the first annular view that correspond to each of the first calibration plate detection boxes as the corresponding first calibration plate images. Each first calibration board image is processed by QR code information recognition to obtain corresponding first QR code information; the first QR code information includes calibration board type information; the calibration board type information includes map information calibration board type and positioning calibration board type; when the calibration board type information is map information calibration board type, the first QR code information is composed of the calibration board type information, map information calibration board identification information, and standard high-precision map information; when the calibration board type information is positioning calibration board type, the first QR code information is composed of the calibration board type information and positioning calibration board identification information; The first target detection box corresponding to the first calibration board image whose calibration board type information is map information calibration board type is used as the corresponding map information calibration board detection box; and the first target detection box corresponding to the first calibration board image whose calibration board type information is positioning calibration board type is used as the corresponding positioning calibration board detection box. The step of generating a corresponding first standard high-precision map by performing standard high-precision map acquisition processing based on the map information calibration board detection box specifically includes: The standard high-precision map information of the first QR code information corresponding to the map information calibration board detection box is taken as the corresponding first standard high-precision map; the map coordinate system of the first standard high-precision map is the road intersection coordinate system; the first standard high-precision map is a semantic map, and the semantic information of the first standard high-precision map includes road semantic information, road edge semantic information, lane semantic information, lane centerline semantic information, lane edgeline semantic information, traffic light semantic information, positioning calibration board semantic information, and pedestrian semantic information; the positioning calibration board semantic information includes positioning calibration board semantic type, positioning calibration board identifier, and positioning calibration board coordinates.

2. The high-precision map processing method according to claim 1, characterized in that, Before generating the corresponding first ring view by taking a surround view of the road ahead during the first vehicle's movement, the method further includes: One map information calibration board and at least two positioning calibration boards are pre-installed on the side of each road leading into the intersection; the map information calibration board and the positioning calibration board are rectangular by default; a QR code containing calibration board type information, map information calibration board identification information and standard high-precision map information is pre-drawn on the map information calibration board; a QR code containing the calibration board type information and positioning calibration board identification information is pre-drawn on the positioning calibration board; the calibration board type information includes map information calibration board type and positioning calibration board type.

3. The high-precision map processing method according to claim 1, characterized in that, The step of confirming whether the road scene ahead is a road intersection scene based on all the first target detection boxes specifically includes: Check whether there is a traffic light type among the target types of all the first target detection boxes obtained; if it is confirmed that there is, then the road scene ahead is confirmed as a road intersection scene; if it is confirmed that there is no, then the road scene ahead is confirmed as a non-road intersection scene.

4. The high-precision map processing method according to claim 1, characterized in that, The step of generating the corresponding first vehicle map coordinates by confirming the vehicle map coordinates based on all the detection frames of the positioning calibration board and the first standard high-precision map specifically includes: The straight-line distance between the current positioning calibration plate detection frame and the workshop is calculated based on the coordinates of the center point of the first detection frame of each of the positioning calibration plate detection frames, and is taken as the corresponding first distance d. The positioning calibration board identifier information of the first QR code information corresponding to each positioning calibration board detection box is used as the corresponding second identifier; and the positioning calibration board coordinates of the positioning calibration board semantic information corresponding to the positioning calibration board identifier that matches each second identifier on the first standard high-precision map are extracted as the corresponding first positioning coordinates. On the first standard high-precision map, a circle is drawn with the first positioning coordinates of each of the positioning calibration board detection boxes as the center and the corresponding first spacing d as the radius to obtain the corresponding first circle; the common intersection point of all the first circles is taken as the corresponding first intersection point; and the number of the first intersection points is counted to generate the corresponding number of first intersection points; the number of the first intersection points is greater than or equal to 1. When the number of the first intersection points is 1, the map coordinates of the first intersection point on the first standard high-precision map are used as the corresponding map coordinates of the first vehicle and output. When the number of first intersection points is greater than 1, each of the first intersection points is traversed. During traversal, the currently traversed first intersection point is taken as the corresponding current intersection point. A corresponding vehicle coordinate system is constructed with the current intersection point as the origin of the vehicle coordinate system and recorded as the corresponding current vehicle coordinate system. The first positioning coordinates of each positioning calibration board detection frame are transformed from the road intersection coordinate system to the current vehicle coordinate system to obtain the corresponding second positioning coordinates. The coordinate deviation between the center point coordinates of the first detection frame and the second positioning coordinates of each positioning calibration board detection frame is calculated to generate the corresponding first coordinate deviation. It is then identified whether the first coordinate deviations corresponding to all the obtained positioning calibration board detection frames are lower than the preset minimum coordinate deviation threshold. If not, the traversal continues to the next first intersection point until the traversal of the last first intersection point ends. If yes, the traversal stops and the map coordinates of the current intersection point on the first standard high-precision map are output as the corresponding first vehicle map coordinates.

5. The high-precision map processing method according to claim 1, characterized in that, The step of performing intersection target fusion processing on the first standard high-precision map based on the first vehicle map coordinates and all the first target detection boxes to generate the corresponding first intersection high-precision map specifically includes: A corresponding vehicle coordinate system is constructed using the first vehicle map coordinates as the origin of the vehicle coordinate system, and denoted as the current vehicle coordinate system. The coordinate transformation relationship between the current vehicle coordinate system and the road intersection coordinate system of the first standard high-precision map is confirmed to obtain the corresponding first transformation relationship. Based on the first transformation relationship, the coordinates of the center point of the first detection box and the orientation angle of the first detection box of each first target detection box are corrected to generate the corresponding second detection box center point coordinates and second detection box orientation angle. On the first standard high-precision map, semantic information of the vehicle is added based on the vehicle orientation, shape and size of the first vehicle and the map coordinates of the first vehicle; and semantic information of the obstacle target is added based on the coordinates of the center point of the second detection box, the orientation angle of the second detection box, the size of the first detection box and the target type of the first detection box of each first target detection box. The first standard high-precision map, after completing the semantic information addition processing for the vehicle and all obstacle targets, is output as the corresponding high-precision map of the first intersection.

6. An apparatus for performing the high-precision map processing method according to any one of claims 1-5, characterized in that, The device includes: a panoramic view processing module, a target detection processing module, a standard high-precision map processing module, and a high-precision map fusion processing module; The surround view processing module is used to take a surround view picture of the road ahead during the first vehicle's driving process to generate a corresponding first surround view. The target detection processing module is used to perform target detection processing on the first ring view to generate multiple first target detection boxes; The standard high-precision map processing module is used to confirm whether the road scene ahead is a road intersection scene based on all the first target detection boxes; if the road scene ahead is confirmed to be a road intersection scene, the map information calibration board detection box and the positioning calibration board detection box are selected from the multiple first target detection boxes; and the standard high-precision map acquisition processing is performed based on the map information calibration board detection box to generate the corresponding first standard high-precision map; The high-precision map fusion processing module is used to perform vehicle map coordinate confirmation processing based on all the positioning calibration board detection boxes and the first standard high-precision map to generate the corresponding first vehicle map coordinates; and to perform intersection target fusion processing on the first standard high-precision map based on the first vehicle map coordinates and all the first target detection boxes to generate the corresponding first intersection high-precision map.

7. An electronic device, characterized in that, include: Memory, processor, and transceiver; The processor is configured to be coupled to the memory, read and execute instructions in the memory to implement the method according to any one of claims 1-5; The transceiver is coupled to the processor, and the processor controls the transceiver to send and receive messages.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a computer, cause the computer to perform the method described in any one of claims 1-5.

Images