A lane line detection method based on laser point cloud

Through the laser point cloud detection method, the problem of large error in lane line detection in high obstacle areas of unmanned ports is solved. Lidar and Kalman filtering technology are used to realize the accurate detection of lane line, improving the safety of port operations and efficiency.

CN118711154BActive Publication Date: 2024-11-05厦门中科星晨科技有限公司
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202411197815.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-29
Publication Date
2024-11-05
Estimated Expiration
2044-08-29

AI Technical Summary

Technical Problem

In unmanned port operations, high obstacle areas cause unstable satellite positioning signals and large lane line detection errors, which may lead to dangerous situations such as unmanned lock-ups or collisions.

Method used

The lane line detection method based on laser point cloud is adopted to obtain the environmental point cloud under the body coordinate system through lidar, obtain the intensity information of the driving area and lane area, generate lane intensity polynomials, update the critical point intensity, perform plane segmentation and linear simulation Together, use Kalman filter to track lane lines to ensure the stability and accuracy of detection.

Benefits of technology

It effectively avoids the error caused by single-frame misdetection, ensures that lane lines can be accurately detected regardless of the vehicle's operation in any scenario, and improves the safety and efficiency of port operations.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118711154B_ABST
    Figure CN118711154B_ABST
Patent Text Reader

Abstract

The present disclosure provides a lane line detection method based on laser point cloud, which relates to the field of autonomous driving technology. In a preset obstacle area, the environment point cloud in the vehicle body coordinate system is obtained by the laser radar of the vehicle, the total intensity information of the driving area corresponding to the driving area and the lane intensity information sequence corresponding to the lane area are obtained, the lane intensity polynomial is generated according to the lane intensity information sequence, and the lane critical point intensity is obtained; the standard intensity coefficient is determined, the lane critical point intensity is updated according to the standard intensity coefficient, and the new lane critical point intensity is obtained; the environment point cloud is plane segmented, the left lane point cloud and the right lane point cloud are screened, and the left lane point cloud and the right lane point cloud are fitted according to the straight line to obtain the target lane line; the target lane line is tracked by using Kalman filtering to obtain the final lane line. Avoid interference of obstacles with height on lane line screening, and ensure that lane detection can be accurately achieved regardless of the vehicle operation in any scenario.
Need to check novelty before this filing date? Find Prior Art

Description

Background Art

[0002] At present, most large-scale transportation relies on sea transportation, which has low prices and high economic benefits. With the development of port freight, the loading and unloading of ports has become unmanned, improving the efficiency and safety of port operations.

[0003] Unmanned container trucks mainly rely on satellite positioning for port transportation operations, but satellite positioning signals are often unstable in some areas of the port. For example, when unmanned container trucks are operating under port yard bridges and quay cranes, weak positioning signals often occur. This is because there is usually a large area of ​​obstruction under the bridge, which causes poor vehicle positioning signals. In this case, unmanned container trucks may go off the track or even collide with other dangerous situations. Traditional solutions use laser lane lines, but due to large areas of high obstacles, the background intensity often cannot be adapted, resulting in unstable detection, large deviations between the detection results and the actual lane lines, and incorrect fitting of lane lines.

[0004] Therefore, in unmanned port operations, the interference of high-altitude obstacles on lane line detection is an urgent problem that needs to be solved.

[0005] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present disclosure, and therefore may include information that does not constitute the prior art known to ordinary technicians in the field. Summary of the invention

[0006] The purpose of the present disclosure is to provide a lane line detection method based on laser point cloud, which at least to a certain extent overcomes the problem of large lane detection error due to related technologies.

[0007] Other features and advantages of the present disclosure will become apparent from the following detailed description, or may be learned in part by the practice of the present disclosure.

[0008] According to one aspect of the present disclosure, a lane line detection method based on laser point cloud is provided, comprising:

[0009] In the preset obstacle area, the vehicle's laser radar is used to obtain the environmental point cloud in the vehicle body coordinate system;

[0010] According to the environmental point cloud, obtaining driving area total intensity information corresponding to the driving area and a lane intensity information sequence corresponding to the lane area, wherein the driving area is smaller than the lane area, the driving area does not include the lane line, and the lane area includes the lane line;

[0011] generating a lane strength polynomial according to the lane strength information sequence, and obtaining a lane critical point strength according to the lane strength polynomial;

[0012] Determine a standard strength coefficient according to the average strength of the total strength information of the driving area, update the lane critical point strength according to the standard strength coefficient, and obtain a new lane critical point strength;

[0013] Performing plane segmentation on the environment point cloud, and screening according to the strength of the new lane critical point to obtain a left lane point cloud and a right lane point cloud;

[0014] Perform straight line fitting according to the left lane point cloud and the right lane point cloud to obtain a target lane line;

[0015] The target lane line is tracked using Kalman filtering to obtain the final lane line.

[0016] In one embodiment of the present disclosure, before obtaining the environment point cloud in the vehicle body coordinate system by using the laser radar of the vehicle in the preset obstacle area, the process includes:

[0017] Performing pre-information verification on the vehicle to determine that the pre-information is not empty, wherein the pre-information includes positioning module information and chassis module information;

[0018] The location information of the vehicle is determined according to the positioning module information, and it is judged whether the area where the vehicle is located belongs to the preset obstacle area.

[0019] In one embodiment of the present disclosure, in the preset obstacle area, obtaining the environment point cloud in the vehicle body coordinate system by using the laser radar of the vehicle includes:

[0020] Obtaining a timestamp of the laser radar;

[0021] If the timestamp of the laser radar is in an updated state, obtaining the environmental laser point through the laser radar;

[0022] The environmental laser points are rotated to the vehicle body coordinate system to obtain the environmental point cloud.

[0023] In one embodiment of the present disclosure, obtaining the driving area total intensity information corresponding to the driving area and the lane intensity information sequence corresponding to the lane area according to the environment point cloud includes:

[0024] Acquire a first point cloud and first intensity information thereof in the driving area;

[0025] Superimposing the first intensity information to obtain total intensity information of the driving area;

[0026] Acquire a second point cloud and second intensity information of the lane area;

[0027] Acquire a third point cloud having a strength less than a preset strength in the second strength information;

[0028] Obtaining third intensity information corresponding to the third point cloud and a point cloud sequence number corresponding to the third intensity information;

[0029] The third intensity information and the point cloud sequence number are stored as the lane intensity information sequence.

[0030] In one embodiment of the present disclosure, generating a lane strength polynomial according to the lane strength information sequence, and obtaining lane critical point strength according to the lane strength polynomial, includes:

[0031] Performing array type conversion on the third intensity information in the lane intensity information sequence and the point cloud sequence number to generate a background intensity polynomial;

[0032] The background intensity polynomial is derived to obtain the lane critical point intensity.

[0033] In one embodiment of the present disclosure, determining a standard strength coefficient according to the average strength of the total strength information of the driving area, updating the lane critical point strength according to the standard strength coefficient, and obtaining a new lane critical point strength includes:

[0034] Obtaining the average intensity of the total intensity information of the driving area;

[0035] Compare the average intensity with a preset threshold value to determine the standard intensity coefficient, wherein the standard intensity coefficient includes a rainy day intensity coefficient and a sunny day intensity coefficient;

[0036] The new lane critical point strength is obtained according to the standard strength coefficient, the average strength and the lane critical point strength.

[0037] In one embodiment of the present disclosure, the plane segmentation of the environment point cloud and the screening according to the strength of the new lane critical point to obtain the left lane point cloud and the right lane point cloud include:

[0038] According to the height information of the environment point cloud, the environment point cloud is plane segmented to obtain a ground point cloud;

[0039] Performing regional judgment on the ground point cloud to determine the left lane regional point cloud and the right lane regional point cloud;

[0040] The left lane area point cloud and the right lane area point cloud are screened according to the strength of the new lane critical point to obtain the left lane point cloud and the right lane point cloud.

[0041] In one embodiment of the present disclosure, the point clouds in the left lane point cloud and the right lane point cloud are traversed to perform nearest edge screening, and after screening, a point cloud close to the inner side is obtained;

[0042] The point clouds in the left lane point cloud and the right lane point cloud are updated according to the point cloud close to the inner side.

[0043] In one embodiment of the present disclosure, performing straight line fitting according to the left lane point cloud and the right lane point cloud to obtain a target lane line includes:

[0044] Linearly segment the left lane point cloud and the right lane point cloud according to a preset straight line normal vector to obtain a straight line point cloud;

[0045] If the number of point clouds of the straight line point cloud exceeds a preset number, re-performing linear segmentation;

[0046] If the number of the straight line point clouds is less than the preset number, obtaining the target straight line point cloud;

[0047] A straight line is fitted according to the target straight line point cloud to obtain the target lane line.

[0048] In one embodiment of the present disclosure, the step of tracking the target lane line using Kalman filtering to obtain a final lane line includes:

[0049] Input the preset sampling time and the corresponding prior information, wherein the prior information includes: chassis speed, moving distance and chassis steering relationship;

[0050] A Kalman filter is performed on the target lane line based on the prior information.

[0051] The embodiment of the present disclosure provides a lane line detection method based on laser point cloud, including: in a preset obstacle area, obtaining an environmental point cloud in a vehicle body coordinate system through a laser radar of a vehicle; according to the environmental point cloud, obtaining the total intensity information of the driving area corresponding to the driving area and the lane intensity information sequence corresponding to the lane area, wherein the driving area is smaller than the lane area, the driving area does not include the lane line, and the lane area includes the lane line, and obtaining the intensity information of the point cloud in different areas. Generate a lane intensity polynomial according to the lane intensity information sequence, and obtain the lane critical point intensity according to the lane intensity polynomial; determine the standard intensity coefficient according to the average intensity of the total intensity information of the driving area, update the lane critical point intensity according to the standard intensity coefficient, and obtain the new lane critical point intensity; perform plane segmentation on the environmental point cloud, and screen according to the new lane critical point intensity to obtain the left lane point cloud and the right lane point cloud; perform straight line fitting according to the left lane point cloud and the right lane point cloud to obtain the target lane line; track the target lane line using Kalman filtering to obtain the final lane line. This embodiment combines the intensity information of the laser point cloud environment to obtain the lane intensity polynomial, and obtains the lane critical point intensity based on the lane intensity polynomial. Kalman filtering is used to track the lane line to solve the problems of unstable detection and single-frame jumps, which can greatly avoid the errors caused by single-frame false detection. Since high obstacles lead to lane line fitting errors, the use of a high-altitude plane segmentation method can avoid interference with lane line screening caused by high-altitude obstacles. Ensure that lane detection can be accurately achieved regardless of the vehicle's operation in any scenario.

[0052] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] The accompanying drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the specification are used to explain the principles of the present disclosure. Obviously, the accompanying drawings described below are only some embodiments of the present disclosure, and for ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without creative work.

[0054] Figure 1 A schematic flow chart of a lane line detection method based on laser point cloud in one embodiment of the present disclosure is shown;

[0055] Figure 2 A schematic diagram of another lane line detection method based on laser point cloud in one embodiment of the present disclosure is shown;

[0056] Figure 3A schematic diagram of a scene showing another lane line detection method based on laser point cloud in an embodiment of the present disclosure;

[0057] Figure 4 A schematic flow chart of a lane line detection method based on laser point cloud in one embodiment of the present disclosure is shown;

[0058] Figure 5 A schematic flow chart of another lane line detection method based on laser point cloud in one embodiment of the present disclosure is shown;

[0059] Figure 6 A schematic flow chart of a lane line detection method based on laser point cloud in another embodiment of the present disclosure is shown;

[0060] Figure 7 A schematic flow chart of a lane line detection method based on laser point cloud in another embodiment of the present disclosure is shown;

[0061] Figure 8 A structural block diagram of a computer device in one embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0062] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that the disclosure will be more comprehensive and complete and to fully convey the concepts of the example embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0063] In addition, the accompanying drawings are only schematic illustrations of the present disclosure and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor devices and / or microcontroller devices.

[0064] Traditional solutions use laser lane lines, but the background intensity often cannot be adapted, resulting in unstable detection, false detection of lane line jumps in a single frame, and incorrect lane line fitting due to large-area obstacles.

[0065] The solution provided by this application, combined with the intensity information of the laser point cloud environment, uses Kalman filtering to track lane lines, which can greatly avoid the errors caused by single-frame false detection and realize unmanned lane detection. For ease of understanding, the following first explains the terms involved in this application.

[0066] Kalman filtering is an algorithm that uses the linear system state equation to optimally estimate the system state through the system input and output observation data. Since the observation data includes the influence of noise and interference in the system, the optimal estimation can also be regarded as a filtering process.

[0067] The solution provided in the embodiment of the present application relates to the lane detection technology for unmanned driving, which is specifically described by the following embodiments:

[0068] Reference Figure 1 As shown, in one embodiment, a lane line detection method based on laser point cloud is provided, comprising:

[0069] S101, in a preset obstacle area, obtaining an environment point cloud in a vehicle body coordinate system by using a laser radar of the vehicle;

[0070] Specifically, the environmental point cloud can contain reflection intensity information (Intensity). The environmental point cloud can be obtained through LiDAR collection. The environmental point cloud reflects the road environment surrounding the vehicle. The preset obstacle area is an area with high obstacles, such as the loading and unloading operation area under the yard crane or the quay crane for unmanned container truck operation. Later, all environmental point clouds can be traversed and the point clouds can be filtered according to the conditions. The detection range of the environmental point cloud is within the preset range of the front, back, left and right of the vehicle, refer to Figure 3 For example, the center of the rear axle of the vehicle body is taken as the origin, the x-axis is directly in front of the vehicle, the y-axis is perpendicular to the vehicle body and points to the left, and the z-axis is perpendicular to the vehicle and points upward (not shown in the figure). The x-axis point cloud point_t.x is less than 25.0m, the y-axis point cloud point_t.y is greater than 0.5m and less than 3.0m, and the y-axis point cloud point_t.y is greater than -3.0m and less than -0.5m. In the area where the z-axis point cloud point_t.z is less than 0.5m, the environmental point cloud is stored in all_point. This area covers all laser point clouds around the road, and covers the environment around the vehicle body more comprehensively. The specific range parameters can be set in the actual scenario and are not limited here.

[0071] Reference Figure 2 As shown, in a specific example, before step S101, the following steps are included:

[0072] S201, verifying the pre-information of the vehicle to determine that the pre-information is not empty, wherein the pre-information includes positioning module information and chassis module information;

[0073] S202, determining the area where the vehicle is located according to the positioning module information, and judging whether the area where the vehicle is located belongs to the preset obstacle area.

[0074] Specifically, a pre-information check is required before starting the detection to check whether the positioning module and chassis module information are empty. If so, the detection will not be started, and the area information is obtained based on the positioning module to determine whether the vehicle is in the preset obstacle area. If it is in the preset obstacle area, the laser lane line detection will be turned on.

[0075] Specifically, the above step S101 includes:

[0076] Obtaining a timestamp of the laser radar;

[0077] Among them, laser radar is used to obtain laser point cloud for lane line detection. When laser lane line detection is turned on, it is checked whether the laser radar has any faults and the verification is performed according to the timestamp.

[0078] If the timestamp of the laser radar is in an updated state, obtaining the environmental laser point through the laser radar;

[0079] Among them, if the timestamp of the laser radar is not updated, it means that the laser radar is abnormally unavailable, and the detection is closed. If the timestamp is updated, it means that the laser radar is normally available, and subsequent detection can be carried out.

[0080] The environmental laser points are rotated to the vehicle body coordinate system to obtain the environmental point cloud.

[0081] The laser radar detects and obtains laser point clouds, which are rotated into the vehicle body coordinate system to facilitate subsequent judgment of the vehicle's surrounding environment.

[0082] S102, acquiring, according to the environmental point cloud, driving area total intensity information corresponding to the driving area and a lane intensity information sequence corresponding to the lane area, wherein the driving area is smaller than the lane area, the driving area does not include a lane line, and the lane area includes the lane line;

[0083] The driving area refers to the center area of ​​the road where the vehicle is about to drive, that is, a clean area without lane lines, for example, Figure 3 The area between the two lane lines A and B in the lane area. The lane area is the road area where you are about to drive. The entire road area includes lane lines A and lane lines B, and the outer edges of its range are dotted lines M and dotted lines N.

[0084] Reference Figure 4 As shown, in a specific example, step S102 includes:

[0085] S401, obtaining a first point cloud and first intensity information thereof within the driving area;

[0086] Specifically, refer to Figure 3 As shown, the center of the rear axle of the vehicle body is taken as the origin. Optionally, the driving area is the area where the x-axis point cloud point_t.x is less than 9.0m and the y-axis point cloud point_t.y is less than 1m and greater than -1, and the first point cloud point_t1 of all laser points in the area is obtained. And the first intensity information point_t.intensity1 of the first laser point cloud.

[0087] S402, superimposing the first intensity information to obtain total intensity information of the driving area;

[0088] The first intensity information point_t.intensity1 is superimposed, that is, the intensity values ​​are added, and the total intensity information of the driving area is obtained and stored in sum_intensity.

[0089] S403, obtaining a second point cloud and second intensity information of the lane area;

[0090] Specifically, combined Figure 3 As shown, with the center of the rear axle of the vehicle body as the origin, the optional lane area is the x-axis point cloud point_t.x less than 9.0m, the y-axis point cloud point_t.y greater than -3.0m and less than 3.0m, and the z-axis point cloud point_t.z less than 0.3m, and the second point cloud point_t2 and the second intensity information point_t.intensity2 of the area are obtained.

[0091] S404, obtaining a third point cloud having a strength less than a preset strength in the second strength information;

[0092] The preset intensity is set based on the statistical results of the background intensity in the center of the road and is used to determine obstacles. Optionally, the preset intensity is set to an intensity value of 150.0. When the second intensity information point_t.intensity2 is less than the intensity information of the point 150.0, it is considered not to be an obstacle and is stored in the intensity array intensity_vec as the third point cloud. If the intensity value is greater than 150.0, the laser point can be considered to be an obstacle and then discarded.

[0093] S405, obtaining third intensity information corresponding to the third point cloud and a point cloud sequence number corresponding to the third intensity information;

[0094] The intensity array intensity_vec is a one-dimensional array, which contains the intensity information of all the third point clouds. It is also necessary to obtain the sequence number of each point cloud in the third point cloud and store it in the intensity sequence sequence_vec.

[0095] S406: Store the third intensity information and the point cloud sequence number as the lane intensity information sequence.

[0096] The lane intensity information sequence includes the third intensity information of the third point cloud and the corresponding point cloud sequence number, that is, the above-mentioned intensity sequence sequence_vec and intensity array intensity_vec.

[0097] S103, generating a lane strength polynomial according to the lane strength information sequence, and obtaining a lane critical point strength according to the lane strength polynomial;

[0098] Specifically, step S103 includes:

[0099] Performing array type conversion on the third intensity information in the lane intensity information sequence and the point cloud sequence number to generate a background intensity polynomial;

[0100] The background intensity polynomial is derived to obtain the lane critical point intensity.

[0101] Specifically, the intensity array intensity_vec is converted and stored in y_vect, and the intensity sequence sequence_vec is converted and stored in x_vect. intensity_vec and sequence_vec are one-dimensional arrays, which are converted into two-dimensional arrays containing the intensity size and point sequence number. There are multiple intensity points in (x_vect, y_vect). For example, an intensity point (x, y) in (x_vect, y_vect), x represents the intensity sequence number, y is the intensity information, and they are arranged in order to get a curve of intensity change. Optionally, any curve can be represented by a 6th-order polynomial: y=A*x^6+B*x^5+C*x^4+D*x^3+E*x^2+F*x+G, where there are A, B, C, D, E, F, G, and 7 unknowns require at least 7 intensity points to solve. By obtaining the intensity points in (x_vect, y_vect), we can get the above unknowns by taking y=A*x^6+B*x^5+C*x^4+D*x^3+E*x^2+F*x+G, and taking the derivative of y to get the background intensity polynomial, which is expressed as:

[0102] fit_p[6]+2.0*fit_p[5]*p+3.0*fit_p[4]*p*p+4.0*fit_p[3]*p*p*p+5.0*fit_p[2]*p*p*p*p+6.0*fit_p[1]*p*p*p*p*p=Kappa.

[0103] Among them, y'=kappa above, the polynomial expression of p is obtained, p is a variable, fit_p[1], fit_p[2], fit_p[3], fit_p[4], fit_p[5], fit_p[6] are known numbers, A, B, C, D, E, F above are equal to fit_p[1], fit_p[2], fit_p[3], fit_p[4], fit_p[5], fit_p[6] respectively, the slope of the inflection point is calculated by derivation of p to obtain the lane critical point strength lane_instensity_down.

[0104] In the above, the intensity information of the laser point cloud environment is combined to obtain the lane intensity polynomial, and the lane critical point intensity is obtained based on the lane intensity polynomial, so that the vehicle surrounding environment can be obtained more accurately.

[0105] S104, determining a standard strength coefficient according to the average strength of the total strength information of the driving area, and updating the lane critical point strength according to the standard strength coefficient to obtain a new lane critical point strength;

[0106] Reference Figure 5 As shown, specifically, step S104 includes:

[0107] S501, obtaining the average intensity of the total intensity information of the driving area;

[0108] Specifically, the total intensity information sum_intensity of the driving area is averaged with the total number of laser point clouds within the range to obtain the average intensity avg_background_intensity.

[0109] S502: Compare the average intensity with a preset threshold value to determine the standard intensity coefficient, wherein the standard intensity coefficient includes a rainy day intensity coefficient and a sunny day intensity coefficient;

[0110] Specifically, the preset threshold can be pre-set according to the current regional conditions. When the average intensity avg_background_intensity is greater than 1.0, the sunny intensity coefficient is obtained as the standard intensity coefficient lane_factor, and the sunny intensity coefficient calibration value is 1.5. When the average intensity avg_background_intensity is less than 1.0, the rainy intensity coefficient is obtained as the standard intensity coefficient lane_factor, and the rainy intensity coefficient calibration value is 6.5.

[0111] S503: Obtain the new lane critical point strength according to the standard strength coefficient, the average strength and the lane critical point strength.

[0112] Specifically, the calculation result of the following expression is obtained:

[0113] lane_intensity_down-avg_background_intensity*lane_factor.

[0114] If the calculated result is greater than 3 and less than 5, the new lane critical point strength is obtained, and its expression is: lane_instensity_down = avg_background_intensity * lane_factor. If the calculated result is not within the range of greater than 3 and less than 5, it is not selected and there is no need to update the lane critical point strength. Further, the background environment around the road where the vehicle is traveling is accurately determined.

[0115] S105, performing plane segmentation on the environment point cloud, and screening according to the strength of the new lane critical point to obtain a left lane point cloud and a right lane point cloud;

[0116] After plane segmentation, the ground point cloud inlanecloud is obtained to accurately filter the lane line point cloud according to the new lane critical point intensity lane_instensity_down. The left and right point clouds will be stored in the left lane point cloud left_lock and the left lane point cloud right_lock. Figure 3 As shown, the left lane point cloud corresponds to area L, and the right lane point cloud corresponds to area R, which distinguishes the left and right lane areas.

[0117] Reference Figure 6 As shown, in a specific example, step S105 includes:

[0118] S601, performing plane segmentation on the environment point cloud according to the height information of the environment point cloud to obtain a ground point cloud;

[0119] Specifically, the segmentation normal vector axis_all is set to segment the point cloud containing height information, perform plane segmentation, obtain the segmented ground point cloud inlanecloud, and separate the ground and non-ground point clouds.

[0120] S602, performing regional judgment on the ground point cloud to determine a left lane regional point cloud and a right lane regional point cloud;

[0121] In the optional example, determine the left lane area information, select the appropriate area point, select the x-axis ground point cloud inlanecloud.points.x greater than -5.0m and less than 25.0m, and the y-axis ground point cloud inlanecloud.points.y greater than 0.5m and less than 2.0m. This area is the left lane area point cloud.

[0122] Determine the right lane area information, select the appropriate area point, select the x-axis ground point cloud inlanecloud.points.x greater than -5.0m and less than 25.0m, the y-axis ground point cloud inlanecloud.points.y greater than -2.0m and less than -0.5m, this area is the right lane area point cloud.

[0123] S603: Filter the left lane area point cloud and the right lane area point cloud according to the strength of the new lane critical point to obtain the left lane point cloud and the right lane point cloud.

[0124] Specifically, the point where the ground point cloud intensity information inlanecloud.points.intensity is greater than lane_instensity_down is stored in the left lane point cloud left_lock; the point where the ground point cloud intensity information inlanecloud.points.intensity is greater than lane_instensity_down is stored in the right lane point cloud right_lock.

[0125] In addition, the method further includes: traversing the point clouds in the left lane point cloud and the right lane point cloud to perform nearest edge screening, and obtaining a point cloud close to the inner side after screening;

[0126] The point clouds in the left lane point cloud and the right lane point cloud are updated according to the point cloud close to the inner side.

[0127] Specifically, the left lane point cloud left_lock and the right lane point cloud right_lock are screened for the nearest edge according to the conditions, and all points in left_lock and right_lock are traversed. The starting point cloud is input first. When the x of the traversed point n When the difference between the value and xx is less than 0.01m and y nWhen the value is less than the minimum miny, the minimum point is entered into test_lock, where n is the serial number of the point cloud. For example, first enter the x0 of the first point, put it into xx, and put the y0 value into miny; the x1 of the next point is less than 0.01m compared with xx, where 0.01m is less than 0.01m, which means getting the point closest to the inside within 1 cm. And y1 is less than 0m compared with miny, then it can be determined that the point (x1, y1) is closer to the inside than the previous point (x0, y0), replace it with xx=x1, miny=y1, traverse all points, get the point closest to the inside in the point cloud, and enter the obtained point cloud near the inside into test_lock. The above screening process makes the screened point set more reliable and accurate, and reduces the number of variables.

[0128] In addition, when traversing the point x n When the difference between the value and xx is greater than 0.01m, the Kalman filter flag lidar_lane_kalman_filter_init_ is judged to be true, and the Kalman filter is turned on to filter near the prediction line, and the currently traversed point is not input into test_lock. If the state is false, the currently traversed point is directly input into test_lock. Copy the test_lock obtained after traversal back to left_lock and right_lock, and update and replace the original value. The starting point cloud refers to the laser point cloud with the first serial number in the left_lock array, which belongs to the point on the inside close to the vehicle position.

[0129] S106, performing straight line fitting according to the left lane point cloud and the right lane point cloud to obtain a target lane line;

[0130] Reference Figure 7 As shown, in a specific example, step S106 includes:

[0131] S701, linearly segmenting the left lane point cloud and the right lane point cloud according to a preset straight line normal vector to obtain a straight line point cloud;

[0132] The left lane point cloud left_lock and the left lane point cloud right_lock are linearly segmented according to the conditions, and the normal vector edge_axis of the straight line is set as the straight line for segmenting the point cloud to obtain the straight line point cloud inlanecloud.

[0133] S702, if the number of point clouds of the straight line point cloud exceeds a preset number, re-performing linear segmentation;

[0134] S703, if the number of the straight line point clouds is less than the preset number, obtaining a target straight line point cloud;

[0135] Specifically, when the number of line point clouds inlanecloud_ exceeds 20, inlanecloud_ is used as input and re-segmented twice to obtain a new inlanecloud_. The preset number can be set in advance based on actual experience to avoid too many points in the line point cloud, which may lead to low fitting accuracy and poor precision. Re-segmentation can improve the accuracy and robustness of fitting.

[0136] S704: Perform straight line fitting according to the target straight line point cloud to obtain the target lane line.

[0137] Perform straight line fitting on inlanecloud_, output the straight line slope param_k and intercept param_b; output the target lane line fitting success flag as true edge_valid. The output straight line is the target lane line, which includes the left lane line and the right lane line. If one of the left and right lane lines is false, restart the above steps and start from step S101. As mentioned above, since the lane line fitting error is caused by the high obstacles, the high plane segmentation method can avoid the interference of the high obstacles on the lane line screening.

[0138] S107, using Kalman filtering to track the target lane line to obtain a final lane line.

[0139] The single-frame jump problem existing in the prior art uses Kalman filtering to track lane lines, which greatly avoids the errors caused by single-frame false detection. A single frame refers to a complete detection output, and a single-frame jump means that there is a large deviation between the detection result and the actual lane line, which will affect the judgment and decision of the downstream module; single-frame false detection also belongs to the category of single-frame jump problems; the use of Kalman filtering can avoid the occurrence of this situation, and will predict the final lane line that should be obtained next time based on the current target lane line.

[0140] In a specific example, step S107 includes:

[0141] Input the preset sampling time and the corresponding prior information, wherein the prior information includes: chassis speed, moving distance and chassis steering relationship;

[0142] A Kalman filter is performed on the target lane line based on the prior information.

[0143] Specifically, the initialization flag is the default value set at the beginning of the program. The initialization flag lidar_lane_kalman_filter_init_=false is obtained. When the fitting success flag of the target lane line is detected to be edge_valid, the initialization flag lidar_lane_kalman_filter_init_=true is assigned to true, and the lane Kalman filter is turned on. Optionally, when lidar_lane_kalman_filter_init_ is false and edge_valid_ is true, the lidar_lane_kalman_filter_init_ state is set to true to turn on the filter.

[0144] Turn on the Kalman filter, input the prior information according to the conditions, input the sampling time with a minimum interval of 0.25s, input the chassis speed, input the moving distance and chassis steering relationship, use the speed, sampling time, moving distance and steering angle to update the state of the Kalman filter and make predictions. Output the slope kalman_k and intercept kalman_b of the predicted straight line, which is the final lane information. The Kalman filter tracks the target lane line to solve the problems of unstable detection and single-frame jumps, which can greatly avoid the errors caused by single-frame false detection. Ensure that lane detection can be accurately achieved regardless of the vehicle's operation in any scenario.

[0145] It will be appreciated by those skilled in the art that various aspects of the present invention may be implemented as a system, method or program product. Therefore, various aspects of the present invention may be specifically implemented in the following forms, namely: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or a combination of hardware and software, which may be collectively referred to herein as a "circuit", "module" or "system".

[0146] Refer to the following Figure 8 An electronic device 800 according to this embodiment of the present invention will be described. Figure 8 The electronic device 800 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0147] like Figure 8 As shown, the electronic device 800 is in the form of a general computing device. The components of the electronic device 800 may include but are not limited to: at least one processing unit 810, at least one storage unit 820, and a bus 830 connecting different system components (including the storage unit 820 and the processing unit 810).

[0148] The storage unit stores program codes, which can be executed by the processing unit 810, so that the processing unit 810 performs the steps according to various exemplary embodiments of the present invention described in the above “Exemplary Method” section of this specification. For example, the processing unit 810 can perform the following steps: Figure 1 The process of lane line detection method based on laser point cloud is shown in .

[0149] The storage unit 820 may include a readable medium in the form of a volatile storage unit, such as a random access storage unit (RAM) 8201 and / or a cache storage unit 8202 , and may further include a read-only storage unit (ROM) 8203 .

[0150] The storage unit 820 may also include a program / utility 8204 having a set (at least one) of program modules 8205, such program modules 8205 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0151] Bus 830 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0152] The electronic device 800 can also communicate with one or more external devices 900 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), and can also communicate with one or more devices that enable users to interact with the electronic device 800, and / or communicate with any device that enables the electronic device 800 to communicate with one or more other computing devices (e.g., routers, modems, etc.). This communication can be carried out through an input / output (I / O) interface 850. In addition, the electronic device 800 can also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs) and / or public networks, such as the Internet) through a network adapter 860. As shown in the figure, the network adapter 860 communicates with other modules of the electronic device 800 through a bus 830. It should be understood that, although not shown in the figure, other hardware and / or software modules can be used in conjunction with the electronic device 800, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, etc.

[0153] Through the description of the above implementation, it is easy for those skilled in the art to understand that the example implementation described here can be implemented by software, or by software combined with necessary hardware. Therefore, the technical solution according to the implementation of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes a number of instructions to enable a computing device (which can be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the implementation of the present disclosure.

[0154] In yet another embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the lane line detection method based on laser point cloud described in any one of the above is implemented.

[0155] In an exemplary embodiment of the present disclosure, a computer-readable storage medium is also provided, on which a program product capable of implementing the above method of the present specification is stored. In some possible implementations, various aspects of the present invention can also be implemented in the form of a program product, which includes a program code, and when the program product is run on a terminal device, the program code is used to enable the terminal device to execute the steps according to various exemplary embodiments of the present invention described in the above "Exemplary Method" section of the present specification.

[0156] A program product for implementing the above method according to an embodiment of the present invention is described, which can adopt a portable compact disk read-only memory (CD-ROM) and include program code, and can be run on a terminal device, such as a personal computer. However, the program product of the present invention is not limited thereto, and in this document, a readable storage medium can be any tangible medium containing or storing a program, which can be used by or in combination with an instruction execution system, an apparatus or a device.

[0157] The program product may be any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination thereof. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0158] Computer readable signal media may include data signals propagated in baseband or as part of a carrier wave, in which readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Readable signal media may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0159] The program code embodied on the readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the foregoing.

[0160] Program code for performing the operations of the present invention may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0161] It should be noted that, although several modules or units of the device for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into multiple modules or units to be embodied.

[0162] In addition, although the steps of the method in the present disclosure are described in a specific order in the drawings, this does not require or imply that the steps must be performed in this specific order, or that all the steps shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps, etc.

[0163] Through the description of the above implementation, it is easy for those skilled in the art to understand that the example implementation described here can be implemented by software, or by combining software with necessary hardware. Therefore, the technical solution according to the implementation of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, a mobile terminal, or a network device, etc.) to execute the method according to the implementation of the present disclosure.

[0164] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. This application is intended to cover any modification, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present disclosure are indicated by the appended claims.

Claims

1. A lane line detection method based on laser point cloud, characterized in that: include: In the preset obstacle area, the vehicle's laser radar is used to obtain the environmental point cloud in the vehicle body coordinate system; According to the environmental point cloud, the total intensity information of the driving area corresponding to the driving area and the lane intensity information sequence corresponding to the lane area are obtained, specifically including: obtaining a first point cloud and its first intensity information in the driving area; superimposing the first intensity information to obtain the total intensity information of the driving area; obtaining a second point cloud and its second intensity information in the lane area; obtaining a third point cloud with less than a preset intensity in the second intensity information; obtaining third intensity information corresponding to the third point cloud, and a point cloud sequence number corresponding to the third intensity information; storing the third intensity information and the point cloud sequence number as the lane intensity information sequence; wherein the driving area is smaller than the lane area, the driving area does not include lane lines, and the lane area includes lane lines; Generate a lane intensity polynomial according to the lane intensity information sequence, and obtain the lane critical point intensity according to the lane intensity polynomial; specifically comprising: performing array type conversion on the third intensity information in the lane intensity information sequence and the point cloud sequence number to generate a background intensity polynomial; Derivative the background intensity polynomial to obtain the lane critical point intensity; Determining a standard strength coefficient according to the average strength of the total strength information of the driving area, updating the lane critical point strength according to the standard strength coefficient, and obtaining a new lane critical point strength; specifically comprising: obtaining the average strength of the total strength information of the driving area; comparing the average strength with a preset threshold to determine the standard strength coefficient, wherein the standard strength coefficient includes a rainy day strength coefficient or a sunny day strength coefficient; obtaining the new lane critical point strength according to the standard strength coefficient, the average strength and the lane critical point strength; Performing plane segmentation on the environment point cloud, and screening according to the strength of the new lane critical point to obtain a left lane point cloud and a right lane point cloud; Perform straight line fitting according to the left lane point cloud and the right lane point cloud to obtain a target lane line; The target lane line is tracked using Kalman filtering to obtain the final lane line.

2. The lane line detection method based on laser point cloud according to claim 1, characterized in that: Before obtaining the environment point cloud in the vehicle body coordinate system by using the laser radar of the vehicle in the preset obstacle area, the method includes: Performing pre-information verification on the vehicle to determine that the pre-information is not empty, wherein the pre-information includes positioning module information and chassis module information; The area where the vehicle is located is determined according to the positioning module information, and it is judged whether the area where the vehicle is located belongs to the preset obstacle area.

3. The lane line detection method based on laser point cloud according to claim 1, characterized in that: The step of obtaining an environment point cloud in a vehicle body coordinate system by using a laser radar of the vehicle in the preset obstacle area includes: Obtaining a timestamp of the laser radar; If the timestamp of the laser radar is in an updated state, obtaining the environmental laser point through the laser radar; The environmental laser points are rotated to the vehicle body coordinate system to obtain the environmental point cloud.

4. The lane line detection method based on laser point cloud according to claim 1, characterized in that: The performing plane segmentation on the environment point cloud and screening according to the strength of the new lane critical point to obtain the left lane point cloud and the right lane point cloud includes: According to the height information of the environment point cloud, the environment point cloud is plane segmented to obtain a ground point cloud; Performing regional judgment on the ground point cloud to determine the left lane regional point cloud and the right lane regional point cloud; The left lane area point cloud and the right lane area point cloud are screened according to the strength of the new lane critical point to obtain the left lane point cloud and the right lane point cloud.

5. The lane line detection method based on laser point cloud according to claim 1, characterized in that: Also includes: Traversing the point clouds in the left lane point cloud and the right lane point cloud to perform nearest edge screening, and obtaining point clouds close to the inner side after screening; The point clouds in the left lane point cloud and the right lane point cloud are updated according to the point cloud close to the inner side.

6. The lane line detection method based on laser point cloud according to claim 1, characterized in that: The performing straight line fitting according to the left lane point cloud and the right lane point cloud to obtain a target lane line includes: Linearly segment the left lane point cloud and the right lane point cloud according to a preset straight line normal vector to obtain a straight line point cloud; If the number of point clouds of the straight line point cloud exceeds a preset number, re-performing linear segmentation; If the number of the straight line point clouds is less than the preset number, obtaining the target straight line point cloud; A straight line is fitted according to the target straight line point cloud to obtain the target lane line.

7. The lane line detection method based on laser point cloud according to claim 1, characterized in that: The method of tracking the target lane line by using Kalman filtering to obtain a final lane line includes: Input the preset sampling time and the corresponding prior information, wherein the prior information includes: chassis speed, moving distance and chassis steering relationship; A Kalman filter is performed on the target lane line based on the prior information.

Citation Information

Patent Citations

  • Interactive extraction method of left and right lane contours based on laser point cloud

    CN109285163A

  • Lane line generation method and device

    CN113835102A

  • Ground point cloud data processing method and device, electronic device and storage medium

    CN116184357A

  • Method and apparatus for detecting lane line

    US20230136798A1