CN-121982669-A - Lane line association method, lane line association device, electronic equipment, mobile tool and readable medium

CN121982669ACN 121982669 ACN121982669 ACN 121982669ACN-121982669-A

Abstract

The application relates to a lane line association method, a lane line association device, an electronic device, a mobile tool and a readable medium, wherein the method comprises the steps of obtaining a first characteristic point set of a lane line of a current frame and a second characteristic point set of a map lane line; the method comprises the steps of respectively projecting a first characteristic point set and a second characteristic point set to an XOZ plane to obtain a first projection point set and a second projection point set, determining a target adjacent point with the shortest distance between each query point in the first projection point set in each second projection point set, calculating a unit normal vector at the position of the target adjacent point, calculating the projection distance between each query point and the target adjacent point on the unit normal vector, calculating the average value of the distances between each query point on the current frame lane line and the map lane line according to each projection distance, taking the average value of the distances as a matching cost value, and determining the map lane line related to each current frame lane line according to each matching cost value. The problem that the lane line association is easy to make mistakes when the deviation angle between the self-vehicle course and the lane line is large is solved.

Inventors

ZHANG XIAOLONG
XIONG HUI

Assignees

北京智行者科技股份有限公司

Dates

Publication Date: 20260505
Application Date: 20251230

Claims (10)

1. A lane-line association method, comprising: Acquiring a first characteristic point set of a current frame lane line and a second characteristic point set of a map lane line corresponding to a current position, wherein the current frame lane line is a lane line in a target frame image acquired at the current position; projecting the first characteristic point set to an XOZ plane under a camera coordinate system to obtain a first projection point set, and projecting the second characteristic point set to the XOZ plane under the camera coordinate system to obtain a second projection point set, wherein the XOZ plane corresponds to a vehicle aerial view angle plane; Determining target adjacent points with shortest distance to each query point in the first projection point set in each second projection point set; Calculating a unit normal vector at the position of the target adjacent point, and calculating the projection distance between the query point and the target adjacent point on the unit normal vector; calculating the average value of the distances from each inquiry point on the current frame lane line to the map lane line according to each projection distance, and taking the average value of the distances as the matching cost value of the current frame lane line and the map lane line; and determining the map lane lines associated with each current frame lane line according to each matching cost value.
2. The method of claim 1, wherein the obtaining a first set of feature points for the current frame lane line and obtaining a second set of feature points for the map lane line corresponding to the current location comprises: Acquiring the target frame image acquired at the current position; Carrying out three-dimensional reconstruction on the current frame lane line by utilizing the target frame image to obtain the first characteristic point set of the current frame lane line, wherein the first characteristic point set comprises three-dimensional coordinates of each characteristic point on the current frame lane line; when the vehicle is at the current position, a map area in a preset range in front of the vehicle body is obtained; extracting each map lane line on the map area, and acquiring the second feature point set of the map lane line, wherein the second feature point set comprises three-dimensional coordinates of each feature point on the map lane line.
3. The method of claim 1, wherein said determining in each of said second set of proxels a target neighbor point that is the shortest distance from each query point in said first set of proxels comprises: Constructing a K-D tree for each second projection point set; And aiming at any query point in the first projection point set, carrying out nearest neighbor searching through the K-D tree to determine the target adjacent point with the shortest distance from the query point on each second projection point set.
4. The method of claim 1, wherein the calculating the projected distance of the query point from the target neighbor point on the unit normal vector comprises: Constructing a target vector by taking the query point as a vector starting point and the target adjacent point as a vector end point; calculating a dot product of the target vector and the unit normal vector; the absolute value of the dot product is determined as the projection distance.
5. The method of claim 1, wherein said determining said map lane line associated with each of said current frame lane lines from each of said matching cost values comprises: integrating the matching cost value of each current frame lane line and each map lane line; based on each matching cost value, constructing a cost matrix taking the map lane line as a row and the current frame lane line as a column; positioning the minimum value of each column in the cost matrix; Comparing the minimum values with a target threshold value, wherein each minimum value corresponds to a group of current frame lane lines and map lane lines; And if the minimum value is smaller than or equal to the target threshold value, judging that the association relationship exists between the current frame lane line corresponding to the minimum value and the map lane line.
6. The method of claim 5, wherein after comparing the minimum value to a target threshold value, the method further comprises: If the minimum value is larger than the target threshold value, judging that the current frame lane line corresponding to the minimum value has no association relationship with the map lane line; and creating a new lane line with the association relation with the lane line of the current frame in the map area.
7. A lane line associating apparatus, comprising: The system comprises an acquisition module, a storage module and a storage module, wherein the acquisition module is used for acquiring a first characteristic point set of a current frame lane line and acquiring a second characteristic point set of a map lane line corresponding to a current position, wherein the current frame lane line is a lane line in a target frame image acquired at the current position; The projection module is used for projecting the first characteristic point set to an XOZ plane under a camera coordinate system to obtain a first projection point set, and projecting the second characteristic point set to the XOZ plane under the camera coordinate system to obtain a second projection point set, wherein the XOZ plane corresponds to a vehicle aerial view angle plane; the first determining module is used for determining target adjacent points with shortest distance to each query point in the first projection point set in each second projection point set; The first calculation module is used for calculating a unit normal vector at the position of the target adjacent point and calculating the projection distance between the query point and the target adjacent point on the unit normal vector; the second calculation module is used for calculating the average value of the distances from each inquiry point on the current frame lane line to the map lane line according to each projection distance, and taking the average value of the distances as the matching cost value of the current frame lane line and the map lane line; and the second determining module is used for determining the map lane lines associated with each current frame lane line according to each matching cost value.
8. An electronic device comprising a memory, a processor, a communication interface and a communication bus, said memory storing a computer program executable on said processor, said memory, said processor communicating with said communication interface via said communication bus, characterized in that said processor, when executing said computer program, implements the steps of the method according to any of the preceding claims 1 to 6.
9. A mobile tool comprising the electronic device of claim 8.
10. A computer readable medium having non-volatile program code executable by a processor, the program code causing the processor to perform the method of any one of claims 1 to 6.

Description

Lane line association method, lane line association device, electronic equipment, mobile tool and readable medium Technical Field The application relates to the technical field of intelligent driving, in particular to a lane line association method, a lane line association device, electronic equipment, a mobile tool and a readable medium. Background The lane lines are used as core information of the driving boundary of the vehicle and are key bases for automatic driving planning and control in a structured environment. In a high-precision map coverage scene, lane line information can be directly obtained from a map, and in a scene without the high-precision map, the lane lines in the environment need to be reconstructed in real time according to the perceived information and the vehicle pose. In the lane line reconstruction process, the data association of the lane line of the current frame and the lane line of the map is a core link. Usually, map lane lines are projected to an XOY plane of a camera coordinate system for association, but only a scene (such as a high-speed main road) with small angle deviation between the heading of a vehicle and the lane lines can be adapted, and under a scene with large angle deviation (such as urban turning, intersection turning, high-speed large-curvature ramp and the like), the map lane lines in the XOY plane are easy to overlap, so that point-to-point matching cost values of the lane lines of the current frame and the lane lines of different maps are converged, an association result error is caused, and lane line reconstruction distortion is caused, so that the application requirements of complex scenes can not be met. That is, lane line association is prone to error when the vehicle heading is at a large angle from the lane line. In view of the above problems, no effective solution has been proposed at present. Disclosure of Invention The application provides a lane line association method, a lane line association device, electronic equipment, a mobile tool and a readable medium, which are used for solving the technical problem that lane line association is easy to make mistakes when the deviation angle between the heading of a self-vehicle and the lane line is large. According to one aspect of the embodiment of the application, the application provides a lane line association method, which comprises the steps of obtaining a first characteristic point set of a current frame lane line and obtaining a second characteristic point set of a map lane line corresponding to a current position, wherein the current frame lane line is a lane line in a target frame image acquired at the current position; the method comprises the steps of projecting a first characteristic point set to an XOZ plane under a camera coordinate system to obtain a first projection point set, projecting a second characteristic point set to the XOZ plane under the camera coordinate system to obtain a second projection point set, wherein the XOZ plane corresponds to a vehicle aerial view angle plane, determining a target adjacent point with the shortest distance to each query point in the first projection point set in each second projection point set, calculating a unit normal vector at the position of the target adjacent point, calculating the projection distance between each query point and the target adjacent point on the unit normal vector, calculating the distance average value between each query point on the current frame lane line and the map lane line according to each projection distance, taking the distance average value as the matching cost value between the current frame lane line and the map lane line, and determining the map lane line associated with each current frame lane line according to each matching cost value. The method comprises the steps of acquiring a first characteristic point set of a current frame lane line, acquiring a second characteristic point set of a map lane line corresponding to a current position, acquiring a target frame image acquired at the current position, carrying out three-dimensional reconstruction on the current frame lane line by utilizing the target frame image to obtain the first characteristic point set of the current frame lane line, wherein the first characteristic point set comprises three-dimensional coordinates of all characteristic points on the current frame lane line, acquiring a map area in a preset range in front of a vehicle body when the vehicle is at the current position, extracting all map lane lines on the map area, and acquiring a second characteristic point set of the map lane line, wherein the second characteristic point set comprises three-dimensional coordinates of all characteristic points on the map lane line. Optionally, determining the target adjacent point with the shortest distance to each query point in the first projection point set in each second projection point set comprises constructing a K-D tree for each second projection p