CN-116778443-B - Lane line detection method and device

CN116778443BCN 116778443 BCN116778443 BCN 116778443BCN-116778443-B

Abstract

The application discloses a lane line detection method and a lane line detection device. The method comprises the steps of obtaining a point cloud map and a target image corresponding to a target environment, inputting the target image into a lane line detection model, detecting lane lines in the target image through the lane line detection model, outputting to obtain a first lane line, extracting ground points from the point cloud map to obtain a ground point cloud, determining a plurality of points meeting preset conditions in the ground point cloud as candidate lane line points, converting the first lane line into the point cloud map to obtain a second lane line, determining the candidate lane line point as a target lane line point when the distance between the candidate lane line point and the second lane line does not exceed a first preset range, and determining the target lane line according to the target lane line point. Thus, the accuracy of lane line detection can be improved.

Inventors

BAO QILIN

Assignees

上海涵润汽车电子有限公司

Dates

Publication Date: 20260508
Application Date: 20230529

Claims (9)

1. A lane line detection method, the method comprising: a point cloud map and a target image corresponding to a target environment are obtained, the target environment comprises lane lines, Inputting the target image into a lane line detection model, detecting lane lines in the target image through the lane line detection model, outputting to obtain a first lane line, Extracting the ground points from the point cloud map to obtain a ground point cloud, Determining a plurality of points meeting preset conditions in the ground point cloud as candidate lane line points, Transforming the first lane line into the point cloud map to obtain a second lane line, Determining the candidate lane line point as a target lane line point under the condition that the distance between the candidate lane line point and the second lane line does not exceed a second preset distance range, Determining a target lane line according to the target lane line point; The method comprises the steps of obtaining multi-frame point cloud data and multiple vehicle pose data of a target environment, carrying out time synchronization on the multi-frame point cloud data and the multiple vehicle pose data to obtain a one-to-one correspondence relationship between the multi-frame point cloud data and the multiple vehicle pose data, determining a transformation matrix for transforming the multiple vehicle pose data into the target vehicle pose data according to the multiple vehicle pose data and the target vehicle pose data, wherein the target vehicle pose data corresponds to target point cloud data, the target point cloud data is the last frame in the multi-frame point cloud data, and transforming the multi-frame point cloud data into a vehicle coordinate system corresponding to the target point cloud data according to the transformation matrix to obtain the point cloud map.
2. The lane line detection method according to claim 1, wherein the point cloud data includes a point cloud time stamp, the vehicle pose data includes a pose time stamp, and the time synchronizing the multi-frame point cloud data and the plurality of vehicle pose data to obtain a one-to-one correspondence between the multi-frame point cloud data and the plurality of vehicle pose data includes: The vehicle pose data corresponding to each frame of point cloud data in the multi-frame point cloud data is determined through the following steps: Determining a first pose timestamp and a second pose timestamp immediately before and after the point cloud timestamp, Determining a target pose timestamp between the first pose timestamp and the second pose timestamp by interpolation, And determining the vehicle pose information corresponding to the target pose timestamp as the vehicle pose information corresponding to the point cloud data.
3. The lane line detection method according to claim 1, wherein before the inputting the target image into the lane line detection model, detecting a lane line in the target image by the lane line detection model, outputting a lane line pixel in the target image, and obtaining a first lane line, the method further comprises: obtaining a calibration internal reference of a camera, wherein the camera is a camera for collecting the target image, And performing de-distortion treatment on the target image according to the calibration internal reference.
4. The lane line detection method according to claim 1, wherein before the inputting the target image into the lane line detection model, detecting a lane line in the target image by the lane line detection model, outputting a lane line pixel in the target image, and obtaining a first lane line, the method further comprises: and carrying out normalization processing on the target image.
5. The lane line detection method according to claim 1, wherein the performing ground point extraction on the point cloud map to obtain a ground point cloud includes: The origin of coordinates of the vehicle coordinate system is taken as the center, a target area is determined in the point cloud map according to a first preset size, Performing grid division on the target area according to a second preset size to obtain a plurality of first grids, Determining differences between maximum and minimum values in the elevation direction of a plurality of points in each first grid, And under the condition that the difference between the maximum value and the minimum value corresponding to the first grid does not exceed the preset difference value, determining the points in the first grid as ground points, and obtaining the ground point cloud.
6. The lane line detection method according to claim 1, wherein the determining that the plurality of points satisfying a preset condition in the ground point cloud are candidate lane line points includes: performing grid division on the ground point cloud according to a third preset size to obtain a plurality of second grids, Determining the sum of the reflected intensities of the plurality of points in each of said second grids, For each row of the second grids, under the condition that the sum of the reflection intensities corresponding to N continuous second grids is sequentially increased, determining a first arrangement sequence number of the last grid in the N second grids, wherein N is a positive integer larger than a preset value, For each row of the second grids, under the condition that the sum of the reflection intensities corresponding to the M second grids is gradually decreased, determining a second arrangement sequence number corresponding to the last grid in the M second grids, wherein M is a positive integer larger than a preset value, And under the condition that the difference value between the first arrangement sequence number and the second arrangement sequence number does not exceed a second preset range, determining the points in a target grid as the candidate lane line points, wherein the target grid comprises grids corresponding to the first arrangement sequence number and the second arrangement sequence number respectively and grids corresponding to the arrangement sequence number between the first arrangement sequence number and the second arrangement sequence number.
7. The lane-line detection method according to claim 1, wherein the transforming the first lane-line into the point cloud map to obtain a second lane-line includes: Transforming the first lane line into a bird's eye view through reverse perspective transformation to obtain a third lane line, And projecting the third lane line into the point cloud map to obtain the second lane line.
8. The lane line detection method according to claim 1, wherein the method further comprises: and determining the attribute information of the target lane line as the attribute information of the second lane line under the condition that the distance between the candidate lane line point and the second lane line does not exceed a first preset range, wherein the attribute information comprises a line type and/or a color.
9. A lane line detection apparatus, characterized in that the apparatus comprises: an acquisition module for acquiring a point cloud map and a target image corresponding to a target environment, wherein the target environment comprises lane lines, An input module for inputting the target image into a lane line detection model, detecting the lane line in the target image by the lane line detection model, outputting to obtain a first lane line, An extraction module for extracting the ground points of the point cloud map to obtain a ground point cloud, A first determining module, configured to determine a plurality of points satisfying a preset condition in the ground point cloud as candidate lane line points, A transformation module for transforming the first lane line into the point cloud map to obtain a second lane line, A second determining module, configured to determine the candidate lane line point as a target lane line point if the distance between the candidate lane line point and the second lane line does not exceed a first preset range, The third determining module is used for determining a target lane line according to the target lane line point; The acquisition module comprises an acquisition sub-module, a synchronization sub-module and a first transformation sub-module, wherein the acquisition sub-module is used for acquiring multi-frame point cloud data and a plurality of vehicle pose data of the target environment, the synchronization sub-module is used for carrying out time synchronization on the multi-frame point cloud data and the plurality of vehicle pose data to obtain a one-to-one correspondence between the multi-frame point cloud data and the plurality of vehicle pose data, the first determination sub-module is used for determining a transformation matrix for transforming the plurality of vehicle pose data into the target vehicle pose data according to the plurality of vehicle pose data and the target vehicle pose data, the target vehicle pose data corresponds to target point cloud data, the target point cloud data is the last frame in the multi-frame point cloud data, and the first transformation sub-module is used for transforming the multi-frame point cloud data into a vehicle coordinate system corresponding to the target point cloud data according to the transformation matrix to obtain the point cloud map.

Description

Lane line detection method and device Technical Field The application belongs to the technical field of data processing, and particularly relates to a lane line detection method and device. Background The automatic driving technology is generally divided into sensing, positioning, decision making, planning and control, wherein a sensing module needs to execute lane line detection, and the detection result can be used for upper-layer functions such as positioning, auxiliary path planning and the like. In addition, the detection result of the lane line is also an indispensable part of the high-precision map. Most of the current lane line detection methods are based on image or point cloud data for detection. The image-based lane line detection generally causes abrasion of the lane line, and the color of the abraded lane line is relatively close to that of the road, so that the lane line detection accuracy is low. Moreover, the detection result generally needs to be converted from two dimensions to three dimensions, which also results in reduced accuracy. Furthermore, lane line detection based on images is relatively affected by light, for example, detection accuracy is also relatively low at night or in and out of tunnels. The lane line detection is carried out based on the point cloud data, and the lane line detection is inaccurate usually due to the problems of lane line breakage, road surface identification, lane line approximation, vehicle shielding, inaccurate road surface segmentation and the like. Therefore, the accuracy of lane line detection based on the image or the point cloud data is low. Disclosure of Invention The embodiment of the application provides a lane line detection method and a lane line detection device, which can improve the accuracy of lane line detection. In a first aspect, an embodiment of the present application provides a lane line detection method, including: A point cloud map and a target image corresponding to a target environment are acquired, the target environment comprises lane lines, Inputting the target image into a lane line detection model, detecting lane lines in the target image through the lane line detection model, outputting to obtain a first lane line, Extracting the ground points from the point cloud map to obtain a ground point cloud, Determining a plurality of points meeting preset conditions in the ground point cloud as candidate lane line points, Transforming the first lane line into a point cloud map to obtain a second lane line, Under the condition that the distance between the candidate lane line point and the second lane line does not exceed the first preset range, determining the candidate lane line point as a target lane line point, And determining a target lane line according to the target lane line point. In a second aspect, an embodiment of the present application provides a lane line detection apparatus, including: An acquisition module for acquiring a point cloud map and a target image corresponding to a target environment, wherein the target environment comprises lane lines, An input module for inputting the target image into a lane line detection model, detecting the lane line in the target image by the lane line detection model, outputting to obtain a first lane line, The extraction module is used for extracting the ground points of the point cloud map to obtain the ground point cloud, A first determining module for determining a plurality of points meeting preset conditions in the ground point cloud as candidate lane line points, A transformation module for transforming the first lane line into the point cloud map to obtain a second lane line, A second determining module for determining the candidate lane line point as the target lane line point in case that the distance between the candidate lane line point and the second lane line does not exceed the first preset range, And the third determining module is used for determining the target lane line according to the target lane line point. In a third aspect, an embodiment of the present application provides an electronic device comprising a processor and a memory storing computer program instructions, The processor, when executing the computer program instructions, implements a lane line detection method as shown in any one of the embodiments of the first aspect. In a fourth aspect, embodiments of the present application provide a computer storage medium having stored thereon computer program instructions which, when executed by a processor, implement the lane line detection method shown in any of the embodiments of the first aspect. In a fifth aspect, embodiments of the present application provide a computer program product, instructions in which, when executed by a processor of an electronic device, cause the electronic device to perform the lane line detection method shown in any one of the embodiments of the first aspect. According to the lane line detection method and device, a point cloud