CN-121978680-A - Road boundary width recognition method, apparatus, radar and readable storage medium

CN121978680ACN 121978680 ACN121978680 ACN 121978680ACN-121978680-A

Abstract

The application relates to a road boundary width recognition method, a device, a radar and a readable storage medium. The method comprises the steps of obtaining point cloud information acquired by a vehicle radar, dividing positions of all stationary point clouds according to a plurality of distance resolution units preset in a plurality of target detection directions of the vehicle radar and transverse positions of all the stationary point clouds, determining target distance resolution units where the peak value of the number of the point clouds in each target detection direction is located, determining single-side road boundary measuring widths of current frames in the corresponding target detection directions according to distance values corresponding to the target distance resolution units for each target detection direction, filtering the single-side road boundary measuring widths of the corresponding current frames according to road width historical information to obtain single-side road boundary widths of the filtered current frames, and detecting false targets in the stationary point clouds based on the single-side road boundary widths of the current frames. The method can improve the detection performance of the radar in a closed scene.

Inventors

WANG XIAOJIE
Hong Shuaixin
WANG MINGHUI

Assignees

福思(杭州)智能科技有限公司

Dates

Publication Date: 20260505
Application Date: 20251128

Claims (10)

1. A method for identifying road boundary width, the method comprising: Determining the road curvature of a vehicle driving road, and acquiring point cloud information acquired by a vehicle radar, wherein the point cloud information comprises all stationary point clouds and the transverse position of each stationary point cloud; Dividing positions of all the stationary point clouds according to a plurality of distance resolution units and transverse positions of each stationary point cloud, which are preset in a plurality of target detection directions of the vehicle radar, and determining a target distance resolution unit in which the peak value of the number of the point clouds in each target detection direction is located from the plurality of distance resolution units; For each target detection direction, determining the unilateral road boundary measurement width of the current frame in the corresponding target detection direction according to the distance value corresponding to the target distance resolution unit; Acquiring road width historical information, and filtering the corresponding unilateral road boundary measuring widths of the current frame according to the road width historical information to obtain the unilateral road boundary width of the current frame after filtering; and detecting a first stationary point cloud distributed in the target detection direction according to the single-side road boundary width, the road curvature and the distance of the current frame, and deleting a false target in the first stationary point cloud.
2. The method according to claim 1, wherein before the dividing of the positions of all the stationary point clouds according to the plurality of distance resolution units preset in the target detection direction of the vehicle radar and the lateral positions of each of the stationary point clouds, the method further comprises: acquiring the turning radius of the vehicle, and determining the curvature radius of the road according to the turning radius; and if the curvature radius is smaller than a preset curvature radius threshold, compensating the transverse positions of all the rest point clouds so that the distribution of the rest point clouds reaches a vertical state, and obtaining the transverse positions after the rest point clouds are compensated.
3. The method according to claim 2, wherein the determining of the preset radius of curvature threshold comprises: Determining a threshold constraint condition, wherein the threshold constraint condition is used for constraining that the transverse positions of at least more than half of stationary point clouds acquired by the vehicle radar belong to the same distance resolution unit under the condition that the transverse position compensation is not performed; And determining the preset curvature radius threshold value based on the maximum longitudinal distance which can be achieved by the vehicle based on the road width, the preset width of the distance resolution unit and the threshold constraint condition.
4. The method according to claim 1, wherein the performing position division on all the stationary point clouds according to a plurality of distance resolution units preset in a plurality of target detection directions of the vehicle radar and lateral positions of each of the stationary point clouds, determining a target distance resolution unit where a peak value of the number of point clouds in each of the target detection directions is located from the plurality of distance resolution units, includes: Dividing positions of all the stationary point clouds according to a plurality of distance resolution units and transverse positions of each stationary point cloud, which are preset in a plurality of target detection directions of the vehicle radar, and determining an intermediate distance resolution unit in which the peak value of the number of the point clouds in each target detection direction is located from the plurality of distance resolution units; for each target detection direction, determining a plurality of candidate distance resolution units comprising the intermediate distance resolution unit from a plurality of the plurality of distance resolution units according to the respective corresponding intermediate distance resolution units; And carrying out region division on the candidate distance resolution units to obtain K sub-distance resolution units, and determining a target distance resolution unit where the peak value of the number of the point clouds is located from the K sub-distance resolution units.
5. The method of claim 4, wherein the obtaining the road width history information, filtering the single-side road boundary measurement widths of the current frame according to the road width history information, and obtaining the single-side road boundary widths of the current frame after filtering, includes: acquiring road width history information; Obtaining the total width of the current frame road after filtering according to the total width of the road of the last frame in the road width history information, preset weights and single-side road boundary measuring widths of the current frame in the target detection direction; Determining a width variation according to the road boundary width of the last frame in the target detection direction, the total width of the filtered current frame, the total width of the road of the last frame and the unilateral road boundary measurement widths of the current frames in a plurality of target detection directions in the road width history information for each target detection direction; And obtaining the single-side road boundary width of the current frame after filtering according to the road boundary width of the previous frame and the width variation.
6. The method of claim 1, wherein the detecting a first stationary point cloud distributed in the target detection direction according to the single-sided road boundary width, the road curvature, and the distance of the filtered current frame, and deleting a false target in the first stationary point cloud, comprises: Determining a theoretical transverse position corresponding to the single-side road boundary width of the filtered current frame according to the single-side road boundary width of the filtered current frame, the road curvature and the distance; The method comprises the steps of obtaining transverse positions of first stationary point clouds distributed in the target detection direction, determining false targets in the first stationary point clouds according to the theoretical transverse positions and the transverse positions of the first stationary point clouds, and deleting the false targets in the first stationary point clouds.
7. The method of any of claims 1 to 6, wherein prior to said deleting a false target in the first stationary point cloud, the method further comprises: obtaining jitter parameters of the total width of a road, a point cloud quantity threshold value and a proportion threshold value; For each target detection direction, determining the first number of the first stationary point clouds distributed in the target detection direction, and determining the ratio of the second number of the stationary point clouds to the total number of the stationary point clouds in the target distance resolution unit; determining a road width confidence according to the jitter parameter, the point cloud quantity threshold, the proportion threshold, the first stationary point cloud quantity and the ratio; and if the confidence coefficient of the road width is larger than the preset confidence coefficient, deleting the false target.
8. A road boundary width recognition apparatus, characterized in that the apparatus comprises: The data determining module is used for determining the road curvature of a vehicle driving road and acquiring point cloud information acquired by a vehicle radar, wherein the point cloud information comprises all stationary point clouds and the transverse position of each stationary point cloud; The position determining module is used for dividing the positions of all the stationary point clouds according to a plurality of distance resolution units preset in a plurality of target detection directions of the vehicle radar and the transverse positions of the stationary point clouds, and determining a target distance resolution unit where the peak value of the number of the point clouds in the target detection directions is located from the plurality of distance resolution units; the road boundary width determining module is used for determining the unilateral road boundary measuring width of the current frame in the corresponding target detection direction according to the distance value corresponding to the target distance resolution unit for each target detection direction; The filtering module is used for acquiring road width historical information, and filtering the corresponding unilateral road boundary measuring widths of the current frame according to the road width historical information to obtain the unilateral road boundary width of the current frame after filtering; And the false target deleting module is used for detecting a first stationary point cloud distributed in the target detection direction according to the single-side road boundary width, the road curvature and the distance of the filtered current frame and deleting false targets in the first stationary point cloud.
9. Radar comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method according to any one of claims 1 to 7 when the computer program is executed.
10. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 7.

Description

Road boundary width recognition method, apparatus, radar and readable storage medium Technical Field The present application relates to the field of radar technologies, and in particular, to a method and apparatus for identifying road boundary width, a radar, and a readable storage medium. Background With the development of radar technology, millimeter wave radar is increasingly used in various fields (e.g., vehicles, ships, etc.). Taking the application to vehicles as an example, along with the intelligent development of automobiles, the millimeter wave radar has the advantages of low cost, no influence of weather, strong anti-interference capability in complex scenes and the like compared with other sensors. However, there are some inherent defects in the target detection performance under certain specific scenes (such as when the vehicle runs in a closed scene of a tunnel lamp), and especially when the radar is in a closed scene such as a tunnel, some multipath and specular ghosts (false targets) are often detected, so that the detection is inaccurate, and the driving function is affected. In order to improve the safety of a vehicle traveling in a closed scene such as a tunnel, a method capable of improving the detection performance of a millimeter wave radar in the closed scene is required. Disclosure of Invention In view of the foregoing, it is desirable to provide a road boundary width recognition method, apparatus, vehicle, computer-readable storage medium, and computer program product that can improve the detection performance of millimeter wave radars in closed scenes. In a first aspect, the present application provides a boundary width identifying method, including: Determining the road curvature of a vehicle driving road, and acquiring point cloud information acquired by a vehicle radar, wherein the point cloud information comprises all stationary point clouds and the transverse position of each stationary point cloud; Dividing positions of all the stationary point clouds according to a plurality of distance resolution units and transverse positions of each stationary point cloud, which are preset in a plurality of target detection directions of the vehicle radar, and determining a target distance resolution unit in which the peak value of the number of the point clouds in each target detection direction is located from the plurality of distance resolution units; For each target detection direction, determining the unilateral road boundary measurement width of the current frame in the corresponding target detection direction according to the distance value corresponding to the target distance resolution unit; Acquiring road width historical information, and filtering the corresponding unilateral road boundary measuring widths of the current frame according to the road width historical information to obtain the unilateral road boundary width of the current frame after filtering; And detecting a first stationary point cloud distributed in the target detection direction according to the single-side road boundary width, the road curvature and the distance of the filtered current frame, and deleting false targets in the first stationary point cloud. In one embodiment, before the position division is performed on all the stationary point clouds according to the plurality of distance resolution units preset in the target detection direction of the vehicle radar and the lateral positions of each of the stationary point clouds, the method further includes: acquiring the turning radius of the vehicle, and determining the curvature radius of the road according to the turning radius; and if the curvature radius is smaller than a preset curvature radius threshold, compensating the transverse positions of all the rest point clouds so that the distribution of the rest point clouds reaches a vertical state, and obtaining the transverse positions after the rest point clouds are compensated. In one embodiment, the determining manner of the preset curvature radius threshold value includes: Determining a threshold constraint condition, wherein the threshold constraint condition is used for constraining that the transverse positions of at least more than half of stationary point clouds acquired by the vehicle radar belong to the same distance resolution unit under the condition that the transverse position compensation is not performed; determining the calculation relation of the preset curvature radius threshold according to the threshold constraint condition; substituting the maximum longitudinal distance which can be achieved by the vehicle based on the road width identification and the preset width of the distance resolution unit into the calculation relation, and determining the preset curvature radius threshold value. In one embodiment, the determining, according to a plurality of distance resolution units preset in a plurality of target detection directions of the vehicle radar and lateral positions of each of