CN-121994257-A - Lane-level vector map generation method and device and electronic equipment

CN121994257ACN 121994257 ACN121994257 ACN 121994257ACN-121994257-A

Abstract

The application discloses a method, a device, electronic equipment and a computer readable storage medium for generating a lane-level vector map, wherein the method comprises the steps of obtaining track data and satellite image data of a target subarea in a target area, obtaining corresponding track flow image data and track speed image data according to the track data of the target subarea, generating target vector map element data of the target subarea according to the track flow image data, the track speed image data and the satellite image data, wherein the target vector map element data comprises target position point characteristic data and target lane marking category data for determining lane markings in the target subarea, and generating the lane-level vector map of the target area according to the target vector map element data. The scheme provided by the application can improve the accuracy of the lane-level vector map generated for the target area.

Inventors

ZHANG DAJI
LIU GUANGWEI
WU ZHENG
ZHANG JIAN

Assignees

高德软件有限公司

Dates

Publication Date: 20260508
Application Date: 20260130

Claims (10)

1. A method of generating a lane-level vector map, the method comprising: Acquiring track data and satellite image data of a target sub-region in a target region; corresponding track flow image data and track speed image data are obtained according to the track data of the target subarea; Generating target vector map element data of the target subarea according to the track flow image data, the track speed image data and the satellite image data, wherein the target vector map element data comprises target position point characteristic data and target lane marking type data for determining lane markings in the target subarea; and generating a lane-level vector map of the target area according to the target vector map element data.
2. The method of claim 1, wherein generating the target vector map element data for the target sub-region from the trajectory flow image data, the trajectory velocity image data, and the satellite imagery data comprises: Respectively carrying out feature extraction on the track flow direction image data, the track speed image data and the satellite image data to obtain track flow direction feature data, track speed feature data and satellite image feature data; Carrying out feature fusion on the track flow direction feature data, the track speed feature data and the satellite image feature data to obtain fusion feature data, and carrying out feature enhancement on the fusion feature data based on a self-attention mechanism to obtain enhancement feature data; and obtaining target vector map element data of the target subarea according to the enhancement characteristic data.
3. The method of claim 2, wherein the feature enhancing the fused feature data based on the self-attention mechanism to obtain enhanced feature data comprises: Position coding is carried out on the fusion characteristic data to obtain fusion characteristic data with position information; inputting the fusion characteristic data with the position information into a standardized pre-layer in a characteristic enhancement encoder, and processing data in the fusion characteristic data through the standardized pre-layer to obtain standardized fusion characteristic data, wherein the data distribution index data of the standardized fusion characteristic data is smaller than a data distribution index data threshold value; And inputting the fusion characteristic data and the standardized fusion characteristic data into a self-attention layer in the characteristic enhancement encoder to obtain enhancement characteristic data.
4. The method according to claim 1, wherein the method further comprises: Acquiring vector priori data of the target subarea, wherein the vector priori data comprises track vector priori data and/or adjacent subarea history vector priori data; the generating the target vector map element data of the target sub-region according to the track flow image data, the track speed image data and the satellite image data comprises the following steps: and generating target vector map element data of the target subarea according to the track flow image data, the track speed image data, the satellite image data and the vector prior data.
5. The method of claim 4, wherein the vector prior data comprises trajectory vector prior data, and wherein the obtaining vector prior data for the target subregion comprises: Sampling target track data with track quality greater than a track quality threshold value from the track data, wherein the target track data comprises position point data; And obtaining track vector priori data corresponding to the target sub-region according to the position point data and the specified lane marking type data, wherein the specified lane marking type data is used for indicating the lane marking corresponding to the target track data as a lane center line.
6. The method of claim 4, wherein the vector prior data comprises contiguous sub-region historical vector prior data, and wherein the obtaining the vector prior data for the target sub-region comprises: acquiring target vector map element data of at least one target adjacent sub-area adjacent to the target sub-area, wherein the target sub-area and the target adjacent sub-area are one of sub-areas of a target area, and two adjacent sub-areas in the target area are overlapped with each other; Acquiring a target overlapping sub-region between the target sub-region and the target adjacent sub-region; and obtaining target overlapping area vector map element data corresponding to the target overlapping sub-area according to the target vector map element data of the target adjacent sub-area, and taking the target overlapping area vector map element data as adjacent sub-area historical vector priori data of the target sub-area.
7. The method according to claim 1, wherein the target location point feature data in the target vector map element data corresponding to each sub-area is target coordinate data of a plurality of location points in the sub-area under a sub-area local coordinate system of the sub-area; the generating the lane-level vector map corresponding to the target area according to the target vector map element data comprises the following steps: For each sub-region, performing coordinate conversion on the target coordinate data according to a coordinate conversion matrix between a sub-region local coordinate system corresponding to the sub-region and a global coordinate system corresponding to the target region to obtain converted target coordinate data of a plurality of position points in the sub-region under the global coordinate system, and taking the converted target coordinate data and target lane marking type data in target vector map element data corresponding to the sub-region as converted target vector map element data corresponding to the sub-region; Determining overlapping position points between a first subarea and a second subarea adjacent to the first subarea and the second subarea in the target area, and carrying out weighting processing on position point characteristic data of the overlapping position points in the first subarea and position point characteristic data of the overlapping position points in the second subarea to obtain overlapping position point characteristic data of the overlapping position points; And generating a lane-level vector map corresponding to the target area according to the overlapping position point characteristic data and the converted target vector map element data.
8. The method according to claim 1, wherein said obtaining trajectory flow image data and trajectory speed image data of said target sub-region from said trajectory data comprises: determining the moving direction and speed of a track point in the track data; Determining target color data corresponding to each track point based on a preset mapping relation between a moving direction and pixel point colors, respectively endowing the corresponding target color data to the pixel points corresponding to each track point, and generating track flow direction image data; And acquiring target brightness data corresponding to each track point based on a preset mapping relation between the speed and the brightness of the pixel points, respectively endowing the corresponding pixel points corresponding to each track point with the corresponding target brightness data, and generating the track speed image data.
9. A lane-level vector map generation apparatus, the apparatus comprising: the acquisition unit is used for acquiring track data and satellite image data of a target subarea in the target area; The obtaining unit is used for obtaining corresponding track flow image data and track speed image data according to the track data of the target subarea; the first generation unit is used for generating target vector map element data of the target subarea according to the track flow image data, the track speed image data and the satellite image data, wherein the target vector map element data comprises target position point characteristic data and target lane marking category data for determining lane markings in the target subarea; And the second generation unit is used for generating a lane-level vector map of the target area according to the target vector map element data.
10. A computer readable storage medium, characterized in that a data processing program is stored, which program is run by a processor, performing the method according to any of claims 1-8.

Description

Lane-level vector map generation method and device and electronic equipment Technical Field The present application relates to the field of computer technologies, and in particular, to a method and apparatus for generating a lane-level vector map, an electronic device, and a computer readable storage medium. Background The intelligent navigation service is typically provided through a standard precision map (Standard Definition Map, also called SD map) in a location services application. However, the SD map is used to provide basic topological connection information of roads, and cannot provide rich road detail information. In contrast, the lane-level vector map can provide rich road detail information such as lane lines, traffic signs and the like, and gradually becomes a key ring in the application of location services. At present, the related technology is to identify ground surface elements such as lane lines, road boundaries and the like through satellite images so as to obtain a lane-level vector map. However, satellite images are limited by shooting angles and resolution, are easy to be interfered by factors such as tree shielding, building projection, vehicle shielding and the like, and key road element information such as lane lines, stop lines and the like is lost or blurred, so that the generated lane-level vector map is low in accuracy. Disclosure of Invention The application provides a method, a device, electronic equipment and a computer readable storage medium for generating a lane-level vector map, which can improve the accuracy of the lane-level vector map. The specific scheme is as follows: In a first aspect, an embodiment of the application provides a method for generating a lane-level vector map, which comprises the steps of obtaining track data and satellite image data of a target subarea in a target area, obtaining corresponding track flow image data and track speed image data according to the track data of the target subarea, generating target vector map element data of the target subarea according to the track flow image data, the track speed image data and the satellite image data, wherein the target vector map element data comprises target position point characteristic data and target lane mark category data for determining a lane mark in the target subarea, and generating the lane-level vector map of the target area according to the target vector map element data. In a second aspect, the embodiment of the application provides a generation device of a lane-level vector map, which comprises an acquisition unit, a first generation unit and a second generation unit, wherein the acquisition unit is used for acquiring track data and satellite image data of a target subarea in a target area, the acquisition unit is used for acquiring corresponding track flow image data and track speed image data according to the track data of the target subarea, the first generation unit is used for generating target vector map element data of the target subarea according to the track flow image data, the track speed image data and the satellite image data, the target vector map element data comprises target position point feature data and target lane mark category data used for determining lane marks in the target subarea, and the second generation unit is used for generating a lane-level vector map of the target area according to the target vector map element data. In a third aspect, the application also provides an electronic device comprising a processor and a memory for storing a data processing program, the electronic device being powered on and executing the program by the processor, to perform the method according to the first aspect. In a fourth aspect, embodiments of the present application also provide a computer readable storage medium storing a data processing program for execution by a processor to perform the method of the first aspect. Compared with the prior art, the application has the following advantages: The method for generating the lane-level vector map comprises the steps of obtaining track data and satellite image data of a target subarea in a target area, obtaining corresponding track flow direction image data and track speed image data according to the track data of the target subarea, generating target vector map element data of the target subarea according to the track flow direction image data, the track speed image data and the satellite image data, wherein the target vector map element data comprises target position point characteristic data and target lane marking type data used for determining lane markings in the target subarea, and generating the lane-level vector map of the target area according to the target vector map element data. Therefore, when the target vector map element data of the target subarea is generated according to the track flow direction image data, the track speed image data and the satellite image data, the track information deletion of the satelli