Road boundary line detection method based on vehicle-mounted millimeter-wave radar

Abstract

The present invention relates to a method for detecting road boundary lines based on vehicle-mounted millimeter-wave radar, comprising: acquiring radar point cloud information and the information of the vehicle body itself of the current road scene; filtering out moving point clouds through static-dynamic separation operation according to the obtained absolute speed of the target; combining the motion state of the current vehicle, and identifying the road scene by applying different point screening and clustering strategies; determining the ROI area to perform the first screening on the static point clouds; using the density-based spatial clustering method to perform clustering processing on the static point clouds after screening to obtain point cloud clusters, and performing the second cluster screening to obtain road boundary clusters; using the least squares method to perform curve fitting on the clustering points that meet the requirements of the second cluster screening, and at the same time performing the third cluster screening to remove some noise clusters and non-road boundary clusters; performing road boundary curve filtering processing on the currently obtained road boundary lines by using the Kalman filtering algorithm to obtain the optimal road curve estimation result of the current frame.

Description

Technical Field

[0001] The present invention relates to the field of intelligent driving, in particular to the field of millimeter-wave radar detection, and specifically refers to a method, system, device, processor and computer-readable storage medium for detecting road boundary lines based on in-vehicle millimeter-wave radar. Background Art

[0002] With the rapid development of the automotive industry and the continuous progress of intelligent technologies, in-vehicle millimeter-wave radar technology has gradually matured and been applied to the field of automotive safety. Millimeter-wave radar has the characteristics of high precision, high resolution, high reliability, etc., and can realize real-time monitoring and high-precision perception of target objects in complex environments.

[0003] Among them, the detection of road boundary lines is crucial for ensuring driving safety. Under adverse weather conditions (such as rain, snow, haze, etc.), the line of sight of the camera is easily affected, while in-vehicle millimeter-wave radar can penetrate weather effects such as rain, snow, and fog to achieve accurate detection of road boundary lines. This helps the driver more accurately judge the road boundary under adverse weather conditions, avoid the vehicle deviating from the road or colliding, thus improving driving safety.

[0004] CN202311552032.8 discloses a method, device, electronic device and storage medium for road boundary detection. Compared with the application that can only detect straight lines using the Hough transform, this technical solution uses the least squares curve fitting method when fitting the road boundary line, and can detect the curved lane boundary line in a non-linear form.

[0005] CN202110528933.8 discloses a method for road boundary detection. Compared with this application, this technical solution can obtain the speed state of the target point cloud, accurately distinguish the static and dynamic states of the target, greatly reduce the number of point cloud targets to be calculated, and in addition, add static point cloud cluster clustering to improve the detection accuracy of the boundary curve.

[0006] CN202210072589.0 discloses a method and system for road boundary detection based on in-vehicle millimeter-wave radar. Compared with this application, this technical solution considers the motion state of the vehicle within the frame period. Especially when the vehicle is turning, the road boundary curve equation in the vehicle's own coordinate system will change. Therefore, the road curve equation after the transformation of the vehicle's own coordinate system is corrected by introducing a rotation matrix and then smoothed and filtered.

[0007] CN202011463411.6 discloses a road boundary detection method based on an in-vehicle millimeter-wave radar. Compared with this application, the curve fitted by the technical solution of the present invention is a quadratic curve, which requires less computational effort compared to the cubic curve fitting mentioned in this application. Moreover, for actual roads, most of them are straight lines or curved boundaries with small curvatures, so there is no need to consider the road fitting of the cubic curve equation.

[0008] Since the above solution does not consider the road curve fitting results of the front and rear frames, there may be a large difference between the results of two frames. Therefore, the technical solution of the present invention introduces a method combining Kalman filtering and matrix transformation, and smooth lane line outputs for each frame can be obtained through the fitting results of the front and rear frames. The technical solution of the present invention also improves the density-based spatial clustering method, and different weights are applied to different regions, different road scenarios, and different clustering directions to reduce the influence of noise points and stationary points of non-road boundaries on the clustering results. Summary of the Invention

[0009] The object of the present invention is to overcome the above-mentioned disadvantages of the prior art and provide a road boundary line detection method, system, device, processor, and computer-readable storage medium based on an in-vehicle millimeter-wave radar.

[0010] In order to achieve the above object, the road boundary line detection method, system, device, processor, and computer-readable storage medium based on an in-vehicle millimeter-wave radar of the present invention are as follows:

[0011] The road boundary line detection method based on an in-vehicle millimeter-wave radar is mainly characterized in that the method includes the following steps:

[0012] (1) The in-vehicle millimeter-wave radar is rigidly connected to the vehicle and installed to obtain the radar point cloud information of the current road scene and the information of the vehicle body itself.

[0013] (2) According to the obtained absolute speed of the target, the moving point cloud is filtered out through the static-dynamic separation operation to determine the absolute speed of the point cloud.

[0014] (3) Combining the motion state of the current vehicle, the road scene is recognized by applying different point screening and clustering strategies.

[0015] (4) Determine the ROI area to perform the first screening on the stationary point cloud.

[0016] (5) Use the density-based spatial clustering method to cluster the stationary point cloud after ROI screening to obtain point cloud clusters, and perform the second cluster screening to obtain road boundary clusters.

[0017] (6) Use the least squares method to perform curve fitting on the clustering points that meet the requirements of the second clustering screening, and at the same time perform the third clustering screening to eliminate some noise clusters and non-road boundary clusters;

[0018] (7) Use the Kalman filtering algorithm to perform road boundary curve filtering on the currently obtained road boundary line, obtain the optimal road curve estimation result of the current frame, and output the corresponding boundary curve.

[0019] Preferably, the step (1) is specifically as follows:

[0020] Specify a Cartesian coordinate system with the centroid of the vehicle body as the coordinate origin, the positive front of the vehicle as the positive x direction of the coordinate system, and the horizontal direction on the positive left side of the vehicle as the positive y direction of the coordinate system. When the on-vehicle millimeter-wave radar detects the target point cloud, through the relative point cloud position, vehicle speed, yaw angular velocity, and installation position information, convert the target position to determine the coordinates in the vehicle body coordinate system. Among them, the conversion method from the point cloud speed to the absolute vehicle speed is as follows:

[0021] ;

[0022] Among them, represents the absolute speed of the target, represents the radial speed measured by the radar, represents the driving speed of the vehicle itself, represents the azimuth angle of the point cloud, represents the yaw angular velocity of the vehicle itself, represents the longitudinal installation position of the radar relative to the vehicle, represents the lateral installation position of the radar relative to the vehicle.

[0023] Preferably, the step (2) is specifically as follows:

[0024] Set the positive speed threshold , when the absolute value of the target absolute speed is less than the threshold , then the point cloud is considered stationary; otherwise, the point cloud is in a moving state, and in this way, filter out the moving point cloud with speed through the static and dynamic separation operation.

[0025] Preferably, the step (3) specifically includes the following steps:

[0026] (3.1) Statistically analyze the height of the absolutely stationary point cloud in the fixed area, determine the ratio Hr of the total number of point clouds that meet the requirements, and if Hr is greater than the empirical parameter Thr, then determine that the current vehicle road scene is a tunnel scene; otherwise, proceed to the next step;

[0027] (3.2) If the current vehicle speed is greater than the first preset speed or the vehicle yaw rate is greater than the first preset angular velocity, it is determined that the current vehicle road scenario is a highway scenario; otherwise, proceed to the next step;

[0028] (3.3) If the current vehicle speed is greater than the second preset speed or the vehicle yaw rate is greater than the second preset angular velocity, it is determined that the current vehicle road scenario is an urban scenario; otherwise, proceed to the next step;

[0029] (3.4) If the current vehicle speed is greater than the third preset speed, it is determined that the current vehicle road scenario is a low-speed scenario; otherwise, proceed to the next step;

[0030] (3.5) If none of the above conditions are met, the current vehicle road scenario is a stationary state.

[0031] Preferably, step (5) specifically includes the following steps:

[0032] (5.1) First, determine the threshold radius and the minimum number of points MinPts. Select an unprocessed data point p from the stationary point cloud and calculate the number of data points within the neighborhood of the data point p.

[0033] (5.2) If the number of data points within the neighborhood of the data point p is greater than or equal to the minimum number of points MinPts, mark the data point p as a core point; if the number of data points within the neighborhood of the data point p is less than MinPts, but this point falls within the neighborhood of other core points, mark the data point p as a boundary point; if the data point p is neither a core point nor a boundary point, mark it as a noise point;

[0034] (5.3) Through each core point, find all its density-reachable data points to form a cluster, and repeat this process continuously until all core points have been processed and all data points p have been marked as belonging to a certain cluster or a noise point;

[0035] (5.4) After clustering is completed, determine whether it is a road boundary cluster by traversing the number of point clouds, the longitudinal length, and the lateral length of each cluster.

[0036] Preferably, step (6) is specifically:

[0037] Use the least squares method to perform curve fitting on the clustering points on the left and right sides of the ego vehicle that are most likely to be road boundary clusters in the following manner:

[0038] ;

[0039] Among them, represents the quadratic coefficient of the fitting curve, represents the linear coefficient of the fitting curve, represents the constant term of the fitting curve, , is the vehicle's own coordinates.

[0040] Preferably, step (6) further includes:

[0041] Calculate the curvature of the fitting result in the following manner, and eliminate the unrealistic road boundary lines according to the yaw rate and curvature :

[0042] ;

[0043] wherein, represents the first derivative of the curve at a certain position, represents the second derivative of the curve at a certain position.

[0044] Preferably, step (7) includes:

[0045] According to the rotation matrix , calculate the prior road curve fitting result of the previous frame after the vehicle turns at the next moment, specifically:

[0046] ;

[0047] wherein, represents the target abscissa after the vehicle turns, represents the target ordinate after the vehicle turns, represents the target abscissa of the previous frame before the vehicle turns, represents the target ordinate of the previous frame before the vehicle turns, represents the angle by which the vehicle rotates within the radar frame period ; ;

[0048] Obtain the new fitting curve coefficients according to the above formula:

[0049] ;

[0050] ;

[0051] ;

[0052] Using the six coefficients obtained from the current frame and the previous frame, obtain a stable coefficient output in the time domain according to the Kalman filter algorithm;

[0053] Obtain the quadratic curve of the road boundary:

[0054] ;

[0055] Among them, is the quadratic coefficient of the road fitting curve after Kalman filter smoothing, is the linear coefficient of the smoothed road fitting curve, is the constant term of the smoothed road fitting curve.

[0056] The road boundary line detection system based on in-vehicle millimeter-wave radar for implementing the above-mentioned method is mainly characterized in that the system includes:

[0057] A radar point cloud preprocessing module, configured to obtain radar point cloud information and vehicle body information of the current road scene;

[0058] A road scene recognition module, connected to the radar point cloud preprocessing module, configured to perform road scene recognition processing by applying different point screening and clustering strategies according to the motion state of the current vehicle;

[0059] A point cloud clustering module, connected to the road scene recognition module, configured to perform clustering processing on the screened point cloud by a spatial clustering method;

[0060] A clustering cluster screening module, connected to the point cloud clustering module, configured to screen out road boundary clusters that meet the requirements from the obtained point cloud;

[0061] A curve fitting module, connected to the clustering cluster screening module, configured to perform curve fitting processing on the clustering points that meet the screening requirements; and

[0062] A road boundary line filtering module, connected to the curve fitting module, configured to perform road boundary curve filtering processing on the currently obtained road boundary line by using the Kalman filter algorithm.

[0063] The road boundary line detection device based on in-vehicle millimeter-wave radar is mainly characterized in that the device includes:

[0064] A processor, configured to execute computer-executable instructions;

[0065] A memory, storing one or more computer-executable instructions, which, when executed by the processor, implement the steps of the above-mentioned road boundary line detection method based on in-vehicle millimeter-wave radar.

[0066] The road boundary line detection processor based on vehicle-mounted millimeter-wave radar is mainly characterized in that the processor is configured to execute computer-executable instructions, and when the computer-executable instructions are executed by the processor, the steps of the above-mentioned road boundary line detection method based on vehicle-mounted millimeter-wave radar are realized.

[0067] The computer-readable storage medium is mainly characterized in that a computer program is stored thereon, and the computer program can be executed by a processor to realize the steps of the above-mentioned road boundary line detection method based on vehicle-mounted millimeter-wave radar.

[0068] By adopting the road boundary line detection method, system, device, processor and computer-readable storage medium of the present invention based on vehicle-mounted millimeter-wave radar, through the point cloud information detected by the vehicle-mounted millimeter-wave radar, the absolute stationary points on the ground after speed conversion calculation are found, the current road scene is estimated in combination with the vehicle speed and the vehicle yaw angular velocity, and then the point clouds that meet the conditions are screened out through ROI to complete the clustering of stationary target clusters and the fitting of road curves, which can effectively eliminate the influence of other stationary obstacles that are not road boundary lines, and compared with the traditional road boundary line detection method, clustering of some stationary point clouds greatly reduces the calculation time of curve fitting. Brief Description of the Drawings

[0069] Figure 1 It is a flowchart of the road boundary line detection method based on vehicle-mounted millimeter-wave radar of the present invention.

[0070] Figure 2 It is a schematic structural diagram of the road boundary line detection system based on vehicle-mounted millimeter-wave radar of the present invention.

[0071] Figure 3 It is an implementation flowchart of the present invention in a specific embodiment.

[0072] Figure 4 It is a schematic diagram of curve fitting in a specific embodiment of the present invention.

[0073] Figure 5 It is a schematic diagram of scene recognition processing in a specific embodiment of the present invention. Detailed Embodiment

[0074] In order to be able to describe the technical content of the present invention more clearly, the following will be further described in combination with specific embodiments.

[0075] Before describing embodiments of the present invention in detail, it should be noted that hereinafter, the terms "comprising", "including" or any other variation thereof are intended to cover non-exclusive inclusion, such that a process, method, article or device comprising a series of elements not only includes those elements but also includes other elements not expressly listed, or elements inherent to such process, method, article or device.

[0076] Please refer to Figure 1 As shown, the method for detecting road boundary lines based on vehicle-mounted millimeter-wave radar, wherein the method comprises the following steps:

[0077] (1) Rigidly connect the vehicle-mounted millimeter-wave radar to the vehicle and install it to obtain the radar point cloud information of the current road scene and the information of the vehicle body itself;

[0078] (2) According to the obtained target absolute speed, filter out the moving point clouds through static-dynamic separation operation to determine the absolute speed of the point clouds;

[0079] (3) Combine the motion state of the current vehicle and perform road scene recognition by applying different point cloud screening and clustering strategies;

[0080] (4) Determine the ROI region to perform the first screening on the static point clouds;

[0081] (5) Use the density-based spatial clustering method to cluster the static point clouds after ROI screening to obtain point cloud clusters, and perform the second cluster screening to obtain road boundary clusters;

[0082] (6) Use the least squares method to perform curve fitting on the clustering points that meet the requirements of the second cluster screening, and at the same time perform the third cluster screening to remove some noise clusters and non-road boundary clusters;

[0083] (7) Use the Kalman filter algorithm to filter the current obtained road boundary line to obtain the optimal road curve estimation result of the current frame and output the corresponding boundary curve.

[0084] As a preferred embodiment of the present invention, step (1) is specifically as follows:

[0085] Designate a Cartesian coordinate system with the centroid of the vehicle body itself as the coordinate origin, the positive front of the vehicle as the positive x direction of the coordinate system, and the horizontal direction to the due left of the vehicle as the positive y direction of the coordinate system. When the vehicle-mounted millimeter-wave radar detects the target point cloud, convert the target position to the vehicle body coordinate system to determine the coordinates through the relative point cloud position, vehicle speed, yaw angular velocity, and installation position information. Among them, the conversion method from the point cloud speed to the absolute vehicle speed is as follows:

[0086] ;

[0087] in, represents the absolute speed of the target, represents the radial velocity measured by the radar, represents the vehicle's speed, Indicates the point cloud orientation angle, represents the yaw angular velocity of the vehicle, Indicates the longitudinal installation position of the radar relative to the vehicle. Indicates the lateral installation position of the radar relative to the vehicle.

[0088] As a preferred embodiment of the present invention, the step (2) is specifically as follows:

[0089] Set the velocity positive threshold , when the absolute value of the target absolute speed Less than threshold , the point cloud is considered to be stationary; otherwise, the point cloud is in motion, and the motion point cloud with speed is filtered out through the static-dynamic separation operation.

[0090] As a preferred embodiment of the present invention, the step (3) specifically includes the following steps:

[0091] (3.1) Count the heights of the absolutely stationary point clouds in a fixed area and determine the ratio Hr of the total number of point clouds that meet the requirements. If Hr is greater than the empirical parameter Thr, the current vehicle road scene is judged to be a tunnel scene. Otherwise, proceed to the next step.

[0092] (3.2) If the current vehicle speed is greater than the first preset speed or the vehicle yaw angular velocity is greater than the first preset angular velocity, the current vehicle road scene is determined to be a high-speed scene, otherwise, proceed to the next step;

[0093] (3.3) If the current vehicle speed is greater than the second preset speed or the vehicle yaw angular velocity is greater than the second preset angular velocity, it is determined that the current vehicle road scene is an urban scene, otherwise, proceed to the next step;

[0094] (3.4) If the current vehicle speed is greater than the third preset speed, the current vehicle road scene is determined to be a low-speed scene, otherwise, proceed to the next step;

[0095] (3.5) None of the above conditions are met, and the current vehicle road scene is stationary.

[0096] As a preferred embodiment of the present invention, the step (5) specifically includes the following steps:

[0097] (5.1) First determine the threshold radius And the minimum number of points MinPts, select an unprocessed data point p from the stationary point cloud, and calculate the data point p in The number of data points in the neighborhood;

[0098] (5.2) If the number of data points p in the neighborhood is greater than or equal to the minimum number of points MinPts, then mark the data point p as a core point; if the number of data points p in the neighborhood is less than MinPts, but this point falls within the neighborhood of other core points, then mark the data point p as a border point; if the data point p is neither a core point nor a border point, then mark it as a noise point;

[0099] (5.3) Through each core point, find all its density-reachable data points to form a cluster, and continuously repeat this process until all core points have been processed and all data points p have been marked as belonging to a certain cluster or a noise point;

[0100] (5.4) After clustering is completed, by traversing the number of point clouds included in each cluster, the longitudinal length and the lateral length of the cluster, thereby determining whether it is a road boundary cluster.

[0101] As a preferred embodiment of the present invention, the step (6) is specifically:

[0102] Use the least squares method to perform curve fitting on the clustering points that are most likely to be road boundary clusters on the left and right sides of the ego vehicle as follows:

[0103] ;

[0104] Wherein, represents the quadratic coefficient of the fitting curve, represents the linear coefficient of the fitting curve, represents the constant term of the fitting curve, , is the ego vehicle coordinate.

[0105] As a preferred embodiment of the present invention, the step (6) further includes:

[0106] Calculate the curvature of the fitting result in the following manner, and eliminate the unrealistic road boundary lines according to the yaw rate and the curvature :

[0107] ;

[0108] Wherein, represents the first derivative of the curve at a certain position, represents the second derivative of the curve at a certain position.

[0109] As a preferred embodiment of the present invention, the step (7) includes:

[0110] According to the rotation matrix , calculate the prior road curve fitting result of the previous frame after the vehicle turns in the next moment, specifically:

[0111] ;

[0112] Among them, represents the target abscissa after the vehicle turns, represents the target ordinate after the vehicle turns, represents the target abscissa of the previous frame before the vehicle turns, represents the target ordinate of the previous frame before the vehicle turns, represents the vehicle within the radar frame period The angle of rotation within; ;

[0113] Obtain new fitting curve coefficients according to the above formula:

[0114] ;

[0115] ;

[0116] ;

[0117] Using the six coefficients obtained from the current frame and the previous frame, obtain a stable coefficient output in the time domain according to the Kalman filter algorithm; ;

[0118] Obtain the road boundary quadratic curve:

[0119] ;

[0120] Among them, is the quadratic term coefficient of the road fitting curve after Kalman filter smoothing, is the linear term coefficient of the smoothed road fitting curve, is the constant term of the smoothed road fitting curve.

[0121] Please refer to Figure 2 shown, the road boundary line detection system based on on-vehicle millimeter-wave radar for implementing the above-described method, wherein the system includes:

[0122] A radar point cloud preprocessing module for obtaining radar point cloud information and self-vehicle body information of the current road scene;

[0123] A road scene recognition module, connected to the radar point cloud preprocessing module, for performing road scene recognition processing by applying different point screening and clustering strategies according to the motion state of the current vehicle;

[0124] A point cloud clustering module, connected to the road scene recognition module, is configured to perform clustering processing on the filtered point cloud through a spatial clustering method;

[0125] A clustering cluster screening module, connected to the point cloud clustering module, is configured to screen out road boundary clusters that meet the requirements from the acquired point cloud;

[0126] A curve fitting module, connected to the clustering cluster screening module, is configured to perform curve fitting processing on the clustering points that meet the screening requirements; and

[0127] A road boundary line filtering module, connected to the curve fitting module, is configured to perform road boundary curve filtering processing on the currently acquired road boundary line by using a Kalman filtering algorithm.

[0128] The following will further elaborate on this technical solution in detail:

[0129] As Figure 1 shown in the flowchart, the present invention proposes a method for detecting road boundary lines based on an in-vehicle millimeter-wave radar. This method mainly involves six modules, including a radar point cloud preprocessing module, a road scene recognition module, a point cloud clustering module, a clustering cluster screening module, a curve fitting module, and a road boundary line filtering module, as Figure 2 shown.

[0130] The in-vehicle millimeter-wave radar is rigidly connected to the vehicle and installed in the front (including the left front, right front, and directly in front), and there are no specific requirements for the vehicle model in this technical solution.

[0131] First, it is necessary to specify a Cartesian coordinate system with the centroid of the vehicle body as the coordinate origin, the directly in front of the vehicle as the positive x direction of the coordinate system, and the horizontal direction directly to the left of the vehicle as the positive y direction of the coordinate system. When the in-vehicle millimeter-wave radar detects the target point cloud, through the relative point cloud position, vehicle speed, yaw angular velocity, and installation position information, the target position can be converted to a determined coordinate in the vehicle body coordinate system, that is, (x, y) is uniquely determined. The conversion formula from point cloud speed to absolute vehicle speed is:

[0132] (1);

[0133] In formula (1), represents the absolute speed of the target, represents the radial speed measured by the radar, represents the driving speed of the vehicle itself, represents the azimuth angle of the point cloud, represents the yaw angular velocity of the vehicle itself, represents the longitudinal installation position of the radar relative to the vehicle, represents the lateral installation position of the radar relative to the vehicle.

[0134] According to the calculated target absolute speed , some moving point clouds with speed can be filtered out through the static-dynamic separation operation. The specific operation is as follows: Set an appropriate positive speed threshold . When the absolute value of the target absolute speed is less than the threshold , the point cloud is considered stationary; otherwise, the point cloud is in a moving state.

[0135] After obtaining the absolute speed of the point cloud and combining the motion state of the vehicle (such as vehicle speed, yaw rate, etc.), the current vehicle driving scenario can be determined as tunnel, highway, urban, low-speed, or vehicle stationary state through the road scene recognition algorithm in Step 3. Different point screening and clustering strategies can be applied in different road scenarios. For example, in the highway scenario, the point screening threshold and the clustering threshold can be appropriately relaxed; in the low-speed and stationary scenarios, the point screening threshold and the clustering threshold are shrunk to improve the accuracy of road boundary fitting. The road scene recognition method is as Figure 3 shown and includes the following main contents:

[0136] Statistically analyze the height of the absolutely stationary point cloud in a fixed area. Usually, the longitudinal range of the area is 15m to 100m, and the transverse range is -6m to 6m. Then calculate the ratio Hr of the number of points with an absolute height greater than 5m and less than 7m to the total number of points with an absolute height less than 20m. If Hr is greater than Thr, where Thr is an empirical parameter, usually 0.21, it can be determined that the current vehicle road scenario is a tunnel scenario. Otherwise, further judge according to the vehicle speed and the vehicle yaw angle: If the vehicle speed is greater than 18m / s or the vehicle speed is greater than 10m / s and the yaw rate is greater than 0.1rad / s, it is judged as a highway scenario; if the vehicle speed is greater than 10m / s or the vehicle speed is 7m / s and the yaw rate is greater than 0.1rad / s, it is judged as an urban scenario; if the vehicle speed is greater than 1m / s, it is judged as a low-speed scenario; if none of the above conditions are met, it is judged that the vehicle is currently in a stationary state.

[0137] Step 4, determine the ROI area (Region of Interest) to screen the stationary point cloud. The number of detections per frame period of the on-vehicle millimeter-wave radar is approximately in the hundreds. Dividing the entire space range where the road boundary curve needs to be detected in the ego-vehicle Cartesian coordinate system, that is, the ROI, can greatly reduce the number of point clouds entering the subsequent boundary line detection, thereby reducing the hardware calculation pressure and improving the real-time performance of the algorithm.

[0138] In step 5, this technical solution uses a density-based spatial clustering method (DBSCAN) to cluster the static point cloud after ROI screening. Its core idea is based on the density of data points. If the density of data points in a region exceeds a certain threshold, these points are divided into a cluster. The specific implementation process is: first determine the threshold radius and the minimum number of points MinPts, then select an unprocessed data point p from the stationary point cloud and calculate p The number of data points in the neighborhood. If the number of data points in the neighborhood is greater than or equal to MinPts, point p is marked as a core point; if the number of data points in the neighborhood is greater than or equal to MinPts, point p is marked as a core point. If the number of data points in the neighborhood is less than MinPts, but the point falls in the neighborhood of other core points, then point p is marked as a boundary point; if point p is neither a core point nor a boundary point, then it is marked as a noise point. Then, through each core point, find all its density-reachable data points to form a cluster. Density reachability is a recursive concept. If point q is within the core point p, then it is marked as a noise point. neighborhood, then q is said to be directly density-reachable from p; if there exists a series of points p1, p2, …, p n , where p i+1 From p i Direct density can be reached, then it is called p n The density can be reached from p1. This process is repeated until all core points are processed and all data points are marked as a cluster or noise point.

[0139] Unlike traditional clustering methods such as partitioning and hierarchical clustering, DBSCAN divides areas with sufficiently high density into clusters and can discover clusters of arbitrary shapes. The algorithm divides data points into three categories: core points, boundary points, and noise points. A core point refers to a point that contains more than the MinPts number of other points in its neighborhood; a boundary point refers to a point that falls within the neighborhood of a core point but does not meet the MinPts number requirement; a noise point is neither a core point nor a boundary point. DBSCAN grows clusters by continuously expanding the neighborhood of core points, which can effectively handle high-dimensional data and outliers. In the clustering process, this technical solution applies different weights to different areas, different road scenes, and different clustering directions to reduce the impact of noise points and non-road boundary stationary points on the clustering results.

[0140] After clustering is completed, the number of point clouds contained in each cluster, the longitudinal length and the lateral length of the cluster are traversed to determine whether it may be a road boundary cluster.

[0141] In step 6, the present invention uses the least square method to perform curve fitting on the cluster points on the left and right sides of the vehicle obtained in step 5 that are most likely to be road boundary clusters. Figure 4As shown, the quadratic curve formula obtained by fitting is as follows:

[0142] (2);

[0143] In formula (2), a represents the quadratic term coefficient of the fitting curve, b represents the linear term coefficient of the fitting curve, and c represents the constant term of the fitting curve.

[0144] In step seven, since the influence of some noise clusters and non-road boundary clusters has been eliminated in step five, and in actual situations, the curvature of a curve is related to different road conditions, the curvature of the fitting result is calculated, and the unrealistic road boundary lines are eliminated according to the yaw rate and curvature.

[0145] The curvature calculation formula is as follows:

[0146] (3);

[0147] In formula (3), represents the first derivative of the curve at a certain position, represents the second derivative of the curve at a certain position.

[0148] In step eight, the present invention uses the Kalman filter algorithm to filter the coefficients of the left and right quadratic curves in the current frame and the coefficients of the left and right quadratic curves in the previous frame to obtain the optimal road curve estimation result in the current frame. In the scenarios of a curve or the self-vehicle turning at a small angle, it will cause errors in the fitting curve equation in the self-vehicle coordinate system. According to the frame period of the radar and the yaw rate of the vehicle, the fitting result of the previous frame curve at the current frame moment is corrected. Therefore, the fitting method is also applicable in the scenarios of a curve or the self-vehicle turning at a small angle.

[0149] According to the rotation matrix: , the prior road curve fitting result of the previous frame after the vehicle turns at the next moment can be calculated. The specific formula is:

[0150] (4);

[0151] In formula (4), represents the target abscissa after the vehicle turns, represents the target ordinate after the vehicle turns, represents the target abscissa of the previous frame before the vehicle turns, represents the target ordinate of the previous frame before the vehicle turns, represents the vehicle rotation angle within the radar frame period , . According to formula (4), it can be calculated that:

[0152] (5);

[0153] (6);

[0154] (7);

[0155] In this step, the current frame and the previous frame are used to obtain Six coefficients can be used to obtain relatively stable coefficient output in the time domain according to the filtering algorithm.

[0156] Take the road boundary line on one side as an example to illustrate the process of Kalman filter processing. Assume that the road boundary curves at time t and time t-1 are:

[0157] (8);

[0158] In the formula, is the corrected fitting coefficient of the vehicle motion at time t-1, is the fitting coefficient at time t. Assume that the state vector , state transfer matrix , the state estimation error covariance matrix is , the noise is , the process noise covariance matrix is , then the prediction equation is:

[0159] (9);

[0160] (10);

[0161] The observation equation is: , is the measurement vector at time t, is the observation matrix, is the residual.

[0162] Kalman Gain :

[0163] (11);

[0164] In the formula, is the measurement noise covariance matrix.

[0165] Calculate the optimal estimate of the posterior state according to the update equation and the posterior state estimation error covariance matrix :

[0166] (12);

[0167] (13);

[0168] Step 9, the obtained quadratic curve of the road boundary is as follows:

[0169] (14);

[0170] In Figure 5 the illustrated embodiment, the obtained vehicle speed carSpeed is 24.94 m / s, the vehicle yaw rate yawRate is -0.01 rad / s. After passing through the road scene recognition module, the calculated Hr is 0.0154, which is less than the threshold 0.21. Therefore, according to Figure 3 the road scene recognition process, it is determined that the current scene is a highway scene. In this embodiment, in combination with the highway scene recognized by the scene recognition, the longitudinal distance of the ROI is set to 0 m to 150 m, the lateral distance is ±15 m, and the height is above 0 m and below 3.5 m. It is set that the number of point clouds in the clustered boundary point cloud cluster is greater than , the length of the clustering cluster is greater than 5 m, and the width of the clustering cluster is less than 0.8 m. , frame period , so the rotation angle , according to Figure 5 the actual sound insulation wall point clouds on both sides of the elevated road, the left road boundary curve of the previous frame is:

[0171] , according to equations (5), (6), and (7), the fitting coefficients after vehicle motion correction within the frame period are calculated as:

[0172] ;

[0173] In this step, using the six coefficients obtained from the current frame and the previous frame, a relatively stable coefficient output in the time domain can be obtained according to the filtering algorithm.

[0174] As calculated above, the left road boundary curve of the previous frame after vehicle motion correction is:

[0175] ;

[0176] Let the measurement noise covariance matrix , and the optimal estimate of the posterior state can be calculated according to equations (11) and (12), that is, the smoothed road boundary curve is:

[0177] ;

[0178] Similarly, performing the same smoothing filter on the right-fitted road boundary curve can obtain a stable lane line output for each frame.

[0179] The road boundary line detection device based on an in-vehicle millimeter-wave radar, wherein the device includes:

[0180] A processor configured to execute computer-executable instructions;

[0181] A memory storing one or more computer-executable instructions, which, when executed by the processor, implement the steps of the above-mentioned road boundary line detection method based on an in-vehicle millimeter-wave radar.

[0182] The road boundary line detection processor based on an in-vehicle millimeter-wave radar, wherein the processor is configured to execute computer-executable instructions, which, when executed by the processor, implement the steps of the above-mentioned road boundary line detection method based on an in-vehicle millimeter-wave radar.

[0183] The computer-readable storage medium, wherein a computer program is stored thereon, and the computer program can be executed by a processor to implement the steps of the above-mentioned road boundary line detection method based on an in-vehicle millimeter-wave radar.

[0184] Any process or method description shown in a flowchart or otherwise described herein can be understood to represent a module, segment, or portion of code including one or more executable instructions for implementing a specific logical function or process. The scope of the preferred embodiments of the present invention includes additional implementations, where functions may be executed in a substantially simultaneous manner or in a reverse order according to the involved functions, rather than in the order shown or discussed, which should be understood by those skilled in the technical field of the embodiments of the present invention.

[0185] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution device.

[0186] Those of ordinary skill in the art of the present technology can understand that all or part of the steps carried by the method of the above embodiments can be completed by instructing relevant hardware through a program. The program can be stored in a computer-readable storage medium, and when the program is executed, it includes one or a combination of the steps of the method embodiments.

[0187] The above-mentioned storage medium can be a read-only memory, a magnetic disk, an optical disc, etc.

[0188] In the description of this specification, the descriptions referring to terms such as "one embodiment", "some embodiments", "example", "specific example", or "embodiment", etc., mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

[0189] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

[0190] By adopting the road boundary line detection method, system, device, processor, and its computer-readable storage medium based on the vehicle-mounted millimeter-wave radar of the present invention, the absolute stationary points on the ground after speed conversion calculation are found through the point cloud information detected by the vehicle-mounted millimeter-wave radar, the current road scene is estimated by combining the vehicle speed and the vehicle yaw angular velocity, and then the point cloud that meets the conditions is screened out through ROI to complete the clustering of the stationary target cluster and the fitting of the road curve. It can effectively eliminate the influence of other stationary obstacles that are not road boundary lines, and compared with the traditional road boundary line detection method, clustering some stationary point clouds greatly reduces the calculation time of curve fitting.

[0191] In this specification, the present invention has been described with reference to its specific embodiments. However, it is obvious that various modifications and transformations can still be made without departing from the spirit and scope of the present invention. Therefore, the specification and the drawings should be regarded as illustrative rather than restrictive.

Claims

1. A method for detecting road boundary lines based on in-vehicle millimeter-wave radar, characterized in that, The method described above includes the following steps: (1) The vehicle-mounted millimeter-wave radar is rigidly connected to the vehicle and installed to obtain the radar point cloud information of the current road scene and the information of the vehicle body itself; (2) According to the obtained absolute speed of the target, the moving point cloud is filtered out through the dynamic and static separation operation to determine the absolute speed of the point cloud; (3) After obtaining the absolute speed of the point cloud, the road scene is recognized in combination with the motion state of the current vehicle, so as to apply different point screening and clustering strategies in different road scenes in the future; (4) Determine the ROI area to perform the first screening on the static point cloud; (5) Use the density-based spatial clustering method to cluster the static point cloud after ROI screening to obtain point cloud clusters, and perform the second cluster screening to obtain road boundary clusters, completing the screening of the radar point cloud; (6) Use the least squares method to perform curve fitting on the clustering points that meet the requirements of the second cluster screening, and at the same time perform the third cluster screening to remove some noise clusters and non-road boundary clusters; (7) Use the Kalman filtering algorithm to perform road boundary curve filtering on the currently obtained road boundary line to obtain the optimal road curve estimation result of the current frame and output the corresponding boundary curve; Among them, the conversion method of the absolute speed of the point cloud is as follows: ； Among them, represents the absolute velocity of the target, represents the radial velocity measured by the radar, represents the driving velocity of the host vehicle, represents the azimuth angle of the point cloud, represents the yaw angular velocity of the host vehicle, represents the longitudinal mounting position of the radar relative to the vehicle, represents the lateral mounting position of the radar relative to the vehicle; The specific steps of step (3) are as follows: (3.1) Statistically analyze the height of the absolutely static point cloud in the fixed area to determine the ratio Hr of the total number of point clouds that meet the requirements. If Hr is greater than the empirical parameter Thr, it is determined that the current vehicle road scene is a tunnel scene; otherwise, proceed to the next step; (3.2) If the current vehicle speed is greater than the first preset speed or the vehicle yaw angular velocity is greater than the first preset angular velocity, it is determined that the current vehicle road scene is a highway scene; otherwise, proceed to the next step; (3.3) If the current vehicle speed is greater than the second preset speed or the vehicle yaw angular velocity is greater than the second preset angular velocity, it is determined that the current vehicle road scene is an urban scene; otherwise, proceed to the next step; (3.4) If the current vehicle speed is greater than the third preset speed, it is determined that the current vehicle road scene is a low-speed scene; otherwise, proceed to the next step; (3.5) If none of the above conditions are met, the current vehicle road scene is in a stationary state.

2. The method for detecting road boundary lines based on in-vehicle millimeter-wave radar according to claim 1, wherein The specific content of step (2) is as follows: Set the positive speed threshold , when the absolute value of the target absolute speed is less than the threshold , it is considered that the point cloud is stationary; otherwise, the point cloud is in a moving state, and in this way, the moving point cloud with speed is filtered out through the static and dynamic separation operation.

3. The method for detecting road boundary lines based on in-vehicle millimeter-wave radar according to claim 1, characterized in that The specific steps of step (5) are as follows: (5.1) First, determine the threshold radius and the minimum number of points MinPts. Randomly select an unprocessed data point p from the static point cloud, and calculate the number of data points within the neighborhood of the data point p for point cloud clustering processing; (5.2) If the number of data points within the neighborhood of data point p is greater than or equal to the minimum number of points MinPts, then mark data point p as a core point; if the number of data points within the neighborhood of data point p is less than MinPts, but this point falls within the neighborhood of other core points, then mark this data point p as a border point; if data point p is neither a core point nor a border point, then mark it as a noise point; If the number of data points within the neighborhood of data point p is greater than or equal to the minimum number of points MinPts, then mark data point p as a core point; if the number of data points within the neighborhood of data point p is less than MinPts, but this point falls within the neighborhood of other core points, then mark this data point p as a border point; if data point p is neither a core point nor a border point, then mark it as a noise point; If the number of data points within the neighborhood of data point p is greater than or equal to the minimum number of points MinPts, then mark data point p as a core point; if the number of data points within the neighborhood of data point p is less than MinPts, but this point falls within the neighborhood of other core points, then mark this data point p as a border point; if data point p is neither a core point nor a border point, then mark it as a noise point; (5.3) Through each core point, find all its density-reachable data points to form a cluster, and repeat this process continuously until all core points are processed and all data points p have been marked as a certain cluster or noise point; (5.4) After the clustering is completed, by traversing the number of point clouds included in each cluster, the longitudinal length and the transverse length of the cluster, determine whether the current cluster is a road boundary cluster, thereby completing the screening of the radar point cloud.

4. The method for detecting road boundary lines based on in-vehicle millimeter-wave radar according to claim 3, characterized in that, The specific content of step (6) is as follows: Use the least squares method to perform curve fitting on the clustering points on the left and right sides of the vehicle itself that are most likely to be road boundary clusters in the following way: ； Among them, represents the quadratic coefficient of the fitting curve, represents the linear coefficient of the fitting curve, represents the constant term of the fitting curve, , is the vehicle coordinate.

5. The method for detecting road boundary lines based on in-vehicle millimeter-wave radar according to claim 3, characterized in that, Step (6) also includes: Calculate the curvature of the fitting result in the following manner, and based on the yaw rate and curvature Eliminate the road boundary lines that do not conform to the reality: ； Among them, represents the first derivative of the curve at a certain position, represents the second derivative of the curve at a certain position.

6. The method for detecting road boundary lines based on in-vehicle millimeter-wave radar according to claim 3, wherein Step (7) includes: According to the rotation matrix , calculate the prior road curve fitting result of the previous frame after the vehicle turns at the next moment, specifically: ； Among them, represents the target abscissa after the vehicle turns, represents the target ordinate after the vehicle turns, represents the target abscissa of the previous frame before the vehicle turns, represents the target ordinate of the previous frame before the vehicle turns, represents the vehicle within the radar frame period and the angle of rotation within, ; Obtain the new fitting curve coefficients according to the above formula: ； ； ； Obtained using the current frame and the previous frame Six coefficients, and based on the Kalman filtering algorithm, a stable coefficient output in the time domain is obtained accordingly; Obtain the road boundary quadratic curve: ； Among them, is the quadratic coefficient of the road fitting curve after Kalman filter smoothing, is the linear coefficient of the smoothed road fitting curve, is the constant term of the smoothed road fitting curve.

7. A road boundary line detection system based on an in-vehicle millimeter-wave radar for implementing the method according to any one of claims 1 to 6, characterized in that, The system described above includes: A radar point cloud preprocessing module, which is used to obtain the radar point cloud information and the information of the vehicle body of the current road scene; A road scene recognition module, connected to the radar point cloud preprocessing module, which is used to perform road scene recognition processing according to the motion state of the current vehicle; A point cloud clustering module, connected to the road scene recognition module, which is used to perform clustering processing on the filtered point cloud by means of spatial clustering; A clustering cluster screening module, connected to the point cloud clustering module, which is used to screen out the road boundary clusters that meet the requirements from the obtained point cloud; A curve fitting module, connected to the clustering cluster screening module, which is used to perform curve fitting processing on the clustering points that meet the screening requirements; and A road boundary line filtering module, connected to the curve fitting module, which is used to perform road boundary curve filtering processing on the currently obtained road boundary line by using the Kalman filtering algorithm.

8. A road boundary line detection device based on an in-vehicle millimeter-wave radar, characterized in that, The device includes: A processor, configured to execute computer-executable instructions; A memory, storing one or more computer-executable instructions, when the computer-executable instructions are executed by the processor, the steps of the road boundary line detection method based on an in-vehicle millimeter-wave radar according to any one of claims 1 to 6 are implemented.

9. A road boundary line detection processor based on an in-vehicle millimeter-wave radar, characterized in that The processor is configured to execute computer-executable instructions, when the computer-executable instructions are executed by the processor, the steps of the road boundary line detection method based on an in-vehicle millimeter-wave radar according to any one of claims 1 to 6 are implemented.

10. A computer-readable storage medium, characterized in that, A computer program is stored thereon, and the computer program can be executed by the processor to implement the steps of the road boundary line detection method based on an in-vehicle millimeter-wave radar according to any one of claims 1 to 6.

Images