CN115015964A - A road edge detection method, device, terminal device and storage medium - Google Patents

Abstract

The present application is applicable to the technical field of radar detection, and provides a road edge detection method, device, terminal device and storage medium. The above-mentioned road edge detection method specifically includes: acquiring point cloud data obtained by scanning the target scene with the laser radar; projecting the scanning points in the point cloud data to a coordinate plane in the radar coordinate system of the laser radar, and obtaining each scan point The projection point of the point in the coordinate plane; the coordinate plane is divided into multiple quadrant areas, and in the same quadrant area, the point cloud radii corresponding to two adjacent scan points on the same scan line of the lidar are calculated. The radius difference of the point cloud corresponding to each scan point is the length of the line segment formed by the projection point of the scan point and the coordinate origin of the coordinate plane; according to the radius difference, determine the road along the target scene from the scan points of the point cloud data Roadside scan points. The embodiments of the present application can avoid determining a false roadside point as a roadside scanning point, and improve the accuracy of roadside point detection.

Description

A road edge detection method, device, terminal device and storage medium

technical field

The present application belongs to the technical field of radar detection, and in particular, relates to a road edge detection method, device, terminal device and storage medium.

Background technique

In autonomous driving solutions, outdoor road edge detection has gradually become a crucial part of perception and positioning. Accurate road edge information can not only provide prior information for vehicle decision planning, but also be used for vehicle positioning.

In the existing road edge detection scheme based on multi-line lidar, as long as the radius difference of two adjacent laser points in the same scan line is large enough, or the gradient difference between the front and rear point clouds in the same scan line and the current point cloud is greater than a certain value threshold, these points are considered to be roadside candidates.

The actual outdoor scenes are more complex, including not only simple scenes such as straight runways and curves, but also complex scenes such as arc-shaped sewers, high dynamic targets, intersections, and T-junctions associated with the roadside. In these complex scenes, the existing methods tend to determine false roadside points such as arc-shaped sewers and tires as roadside points, and the detection accuracy of roadside points is low.

SUMMARY OF THE INVENTION

Embodiments of the present application provide a road edge detection method, device, terminal device, and storage medium, which can avoid determining a false road edge point as a road edge scan point, and improve the accuracy of road edge point detection.

A first aspect of the embodiments of the present application provides a road edge detection method, including:

Obtain point cloud data obtained by scanning the target scene with lidar;

Projecting the scanning points in the point cloud data to a coordinate plane in the radar coordinate system of the lidar to obtain a projection point of each of the scanning points in the coordinate plane;

Divide the coordinate plane into a plurality of quadrant areas, and in the same quadrant area, calculate the distance between the point cloud radii corresponding to the two adjacent scan points on the same scan line of the lidar. Radius difference, the point cloud radius corresponding to each of the scanning points is the length of the line segment formed by the projection point of the scanning point and the coordinate origin of the coordinate plane;

According to the radius difference, a roadside scan point of a roadside in the target scene is determined from the scan points of the point cloud data.

In some implementations of the first aspect, in the same quadrant area, the distance between the point cloud radii corresponding to two adjacent scan points on the same scan line of the lidar is calculated. Before the radius difference, the road edge detection method includes: calculating, on the same rotational azimuth of the lidar, the connecting line between the two scanning points respectively located on two adjacent scanning lines of the lidar and the The included angle between the coordinate planes; according to the angle of the included angle, the ground point is determined from the scanning points of the point cloud data; the calculation is performed in the same scan of the lidar in the same quadrant area The radius difference between the point cloud radii corresponding to the two adjacent scanning points on the line, including: in the same quadrant area, calculating the two adjacent scanning points on the same scanning line of the lidar. The radius difference between the point cloud radii corresponding to the ground points respectively.

In some implementations of the first aspect, after the ground point is determined from the scanning points of the point cloud data according to the angle of the included angle, the road edge detection method further includes: acquiring the laser light The first distance between the radar and the road in the target scene; according to the first distance, determine the radar projection point of the lidar on the road; according to the radar projection point and the location of each ground point The coordinate position in the radar coordinate system is selected, and the ground points whose distance from the radar projection point is greater than the distance threshold are screened out.

In some implementations of the first aspect, in the same quadrant area, the radius between the point cloud radii corresponding to the two adjacent scan points on the same scan line of the lidar is calculated The difference includes: sequentially taking each of the scan points as the current scan point, and calculating the radius difference between the point cloud radius corresponding to the current scan point and the point cloud radius corresponding to the next scan point, wherein the next scan point The scan point is the next scan point on the same scan line in the same quadrant area according to the rotation direction of the lidar.

In some implementations of the first aspect, the determining, according to the radius difference, a roadside scan point of a roadside in the target scene from the scan points of the point cloud data includes: if the The radius difference between the point cloud radius corresponding to the current scan point and the point cloud radius corresponding to the next scan point satisfies the difference range corresponding to the quadrant area where the current scan point is located, then the current scan point is used as the The scanning point along the road; or, if the radius difference between the point cloud radius corresponding to the current scanning point and the point cloud radius corresponding to the next scanning point satisfies the difference value corresponding to the quadrant area where the current scanning point is located range, take the current scan point as the roadside scan point as the first roadside candidate point; sequentially calculate every two adjacent first roadside candidate points on the same scan line of the lidar to form The slope of the straight line; according to the slope, the roadside scanning point is determined from the first roadside candidate points.

In some implementations of the first aspect, the determining the roadside scan point from the first roadside candidate points according to the slope includes: if there are N consecutive first roadside scan points on the same scan line If the slope difference between the slopes corresponding to the candidate roadside points is less than the difference threshold, the consecutive N first candidate roadside points are used as the roadside scanning points, where N is greater than 2; or, If the slope difference between the slopes corresponding to the N consecutive first roadside candidate points on the same scan line is smaller than the difference threshold, the consecutive N first roadside candidate points are used as the second candidate point; in each quadrant area in turn, determine the second candidate point on each scan line that is closest to the lidar distance; determine the roadside scan point according to the determined second candidate point .

In some implementations of the first aspect, the determining the roadside scan point according to the determined second candidate point includes: using the determined second candidate point as the roadside scan point; Or, take the determined second candidate point as a third candidate point; perform straight line fitting on the third candidate point in each of the quadrant regions, and fit the third candidate point according to the straight line obtained by fitting The non-outlier points among the three candidate points are used as the roadside scanning points.

A road edge detection device provided by the second aspect of the embodiment of the present application includes:

The point cloud data acquisition unit is used to acquire the point cloud data obtained by scanning the target scene with the lidar;

a scanning point projection unit, configured to project the scanning points in the point cloud data to a coordinate plane in the radar coordinate system of the lidar to obtain a projection point of each of the scanning points in the coordinate plane;

a radius difference calculation unit, configured to divide the coordinate plane into a plurality of quadrant areas, and in the same quadrant area, calculate the corresponding corresponding scan points of the two adjacent scan points on the same scan line of the lidar respectively The radius difference between the point cloud radii, the point cloud radius corresponding to each scan point is the length of the line segment formed by the projection point of the scan point and the coordinate origin of the coordinate plane;

A roadside point determination unit, configured to determine a roadside scan point of a roadside in the target scene from the scan points of the point cloud data according to the radius difference.

A third aspect of an embodiment of the present application provides a terminal device, including a memory, a processor, and a computer program stored in the memory and running on the processor, where the processor implements the above when executing the computer program The steps of the road edge detection method.

A fourth aspect of the embodiments of the present application provides a computer-readable storage medium, where the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, implements the steps of the above-mentioned road edge detection method.

A fifth aspect of the embodiments of the present application provides a computer program product that, when the computer program product runs on a terminal device, enables the terminal device to execute the road edge detection method described in any one of the first aspects above.

In the embodiment of the present application, the point cloud data obtained by scanning the target scene with the laser radar is obtained, and the scanning points in the point cloud data are projected to a coordinate plane in the radar coordinate system of the laser radar, so that each scan point is obtained. The projection point of the point in the coordinate plane, and then divide the coordinate plane into multiple quadrant areas, and in the same quadrant area, calculate the difference between the point cloud radii corresponding to the two adjacent scan points on the same scan line of the lidar. The radius difference between the scan points, the point cloud radius corresponding to each scan point is the length of the line segment formed by the projection point of the scan point and the coordinate origin of the coordinate plane. The radius difference has different positive and negative relationships in different quadrants. According to the radius difference, the terminal device can determine the roadside scanning points of the roadside in the target scene from the scanning points of the point cloud data, so as to avoid the false roadside points being judged as The roadside scanning point improves the accuracy of roadside point detection.

Description of drawings

In order to illustrate the technical solutions in the embodiments of the present application more clearly, the following briefly introduces the accompanying drawings that need to be used in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only for the present application. In some embodiments, for those of ordinary skill in the art, other drawings can also be obtained according to these drawings without any creative effort.

1 is a scanning diagram obtained by scanning a multi-line laser radar provided by an embodiment of the present application in a scene including an arc-shaped sewer;

2 is a scanning diagram obtained by scanning a multi-line laser radar provided by an embodiment of the present application in a scene where a dynamic target appears at a T-shaped intersection or a crossroad intersection;

3 is a schematic diagram of an implementation flow of a road edge detection method provided by an embodiment of the present application;

FIG. 4 is a schematic flowchart of a specific process for removing point clouds with inconspicuous features in ground points provided by an embodiment of the present application;

FIG. 5 is a schematic flowchart of a specific flow of step S304 provided by an embodiment of the present application;

FIG. 6 is a schematic flowchart of a specific flow of step S503 provided by an embodiment of the present application;

FIG. 7 is a schematic flowchart of a specific flow of step S603 provided by an embodiment of the present application;

8 is a schematic diagram of a specific flow of road edge detection provided by an embodiment of the present application;

9 is a schematic structural diagram of a road edge detection device provided by an embodiment of the present application;

FIG. 10 is a schematic structural diagram of a terminal device provided by an embodiment of the present application.

Detailed ways

In order to make the purpose, technical solutions and advantages of the present application more clearly understood, the present application will be described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, but not to limit the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative work belong to the protection of the present application.

In autonomous driving solutions, outdoor road edge detection has gradually become a crucial part of perception and positioning. Accurate road edge information can not only provide prior information for vehicle decision planning, but also be used for vehicle positioning.

In the existing road edge detection scheme based on multi-line lidar, as long as the radius difference of two adjacent laser points in the same scan line is large enough, or the gradient difference between the front and rear point clouds in the same scan line and the current point cloud is greater than a certain value threshold, these points are considered to be roadside candidates.

The actual outdoor scenes are more complex, including not only simple scenes such as straight runways and curves, but also complex scenes such as arc-shaped sewers, high dynamic targets, intersections, and T-junctions associated with the roadside.

Please refer to Figure 1 and Figure 2. Among them, Fig. 1 shows a scanning diagram obtained by scanning a multi-line lidar in a scene containing an arc-shaped sewer. The curved sewer is the pavement used for drainage between the road and the actual curb. This part of the pavement also belongs to the passable area, not the curb. However, in the existing method, due to a certain radian between the curved sewer and the road, two adjacent scanning points in the same scan line (one laser point is the road scanning point, the other is the scanning point of the curved sewer). ), the difference in point cloud radius is large, and the terminal device will determine the scanning point of the arc-shaped sewer as the scanning point along the road.

Figure 2 shows the scanning diagram obtained by multi-line lidar scanning in a scene where dynamic targets appear at a T-shaped intersection or a crossroad intersection. Assuming that the dynamic target is a vehicle, it should be understood that when the terminal device is located at the center of a T-shaped intersection or intersection, the surrounding points do not include road edge points. Due to the sudden appearance of vehicles on the road, the point cloud radius difference between two adjacent scan points in the same scan line (one of the laser points is the road scan point and the other is the tire scan point) is quite different. At this time, the terminal equipment The scan point of the tire will be determined as the road edge scan point.

Therefore, in these complex scenarios, the false roadside points are often determined as roadside points by the existing methods, and the detection accuracy of the roadside points is low, thereby reducing the reliability of subsequent vehicle decision-making planning.

In view of this, the present application proposes a road edge detection method to solve the above problems.

In order to illustrate the technical solutions of the present application, the following specific embodiments are used for description.

FIG. 3 shows a schematic diagram of the implementation flow of a road edge detection method provided by an embodiment of the present application. The method can be applied to a terminal device, and can be applied to avoid false road edge points from being judged as road edge scanning points, and improve road edge detection. Accuracy of edge detection.

Wherein, the above-mentioned terminal device may be a vehicle, a roadside unit, a robot, etc. equipped with a radar. In this case, the terminal device may use its own radar to detect roadside points. The above-mentioned terminal device may also be an intelligent terminal such as a computer and a mobile phone. In this case, the terminal device can obtain information collected by radar of other devices to detect roadside points.

Specifically, the above road edge detection method may include the following steps S301 to S304.

Step S301, acquiring point cloud data obtained by scanning the target scene with the lidar.

In the embodiments of the present application, the above-mentioned lidar may specifically refer to a multi-line lidar, and a multi-line lidar is a lidar including at least two laser beams. The multi-line lidar uses the distribution of multiple laser emitters in the vertical direction. After one scan, the second scan is performed by the rotation of the motor, and finally multiple scan lines can be formed. The number of laser beams of a multi-line lidar determines the number of scan lines. The rotation angle of the multi-line laser radar determines the sensing range of the laser radar. In the embodiment of the present application, the radar can rotate 360°.

Exemplarily, if the terminal device is a vehicle, the multi-line lidar can be installed on the roof or front of the vehicle.

In the embodiment of the present application, the target scene is a scene scanned by a lidar. Point cloud data refers to a set of points consisting of scan points containing three-dimensional coordinates, which can represent the distance between each point in the target scene and the lidar.

Wherein, the three-dimensional spatial geometric position information of each scanning point can be represented by coordinates (x, y, z) in the radar coordinate system. Among them, the radar coordinate system is usually a three-dimensional coordinate system established with the position of the radar as the origin and the plane parallel to the bottom surface of the radar installation as the xoy plane.

In some specific scenarios of this application, the point cloud data collected by the lidar may be the point cloud data in the road scene where the vehicle is located, including the points of objects such as trees, fences, roads, road edges, and false road edges on the road. cloud data.

Step S302, project the scanning points in the point cloud data to a coordinate plane in the radar coordinate system of the lidar to obtain the projection point of each scanning point in the coordinate plane.

In the embodiment of the present application, the scanning point, that is, the laser point, based on the three-dimensional coordinates of each scanning point in the radar coordinate system, the scanning point can be projected to any coordinate plane in the radar coordinate system of the lidar, and each scanning point can be obtained. The projection point of each scan point in the coordinate plane.

Since the xoy surface of the radar coordinate system is usually parallel to the radar installation bottom surface, and the radar installation bottom surface is usually parallel to the ground, the terminal equipment can project the scanning point to the xoy surface of the radar coordinate system, which is convenient for road edge detection.

Step S303: Divide the coordinate plane into multiple quadrant regions, and in the same quadrant region, calculate the radius difference between the point cloud radii corresponding to two adjacent scanning points on the same scanning line of the lidar.

Among them, the quadrant area is also the four quadrants obtained by dividing the origin and the coordinate axis of the coordinate plane. After projecting the scanning points on the coordinate plane, the terminal device can obtain the two-dimensional coordinates of the projected points of each scanning point in the coordinate plane, and based on the two-dimensional coordinates, the quadrant area where each projected point is located can be determined.

In some embodiments of the present application, each scan line has a unique identifier, such as a line number of the scan line. Scan points on the same scan line have the same line number, but different scan numbers. Based on the line number and scan sequence number of each scan point obtained during the radar scanning process, two adjacent scan points in the same quadrant area and on the same scan line can be determined.

More specifically, the point cloud data obtained by multi-line lidar scanning can be represented in the form of a matrix, each row is a scan line, each column is a scan line, and the points on two adjacent columns in the same row in the matrix are the same. Two adjacent scan points on a scan line.

At this point, the terminal device can calculate the radius difference between the point cloud radii corresponding to two adjacent scanning points on the same scanning line of the lidar in the same quadrant area. The point cloud radius corresponding to each scan point is the length of the line segment formed by the projection point of the scan point and the coordinate origin of the coordinate plane.

Specifically, the terminal device may sequentially take each scan point as the current scan point, and calculate the radius difference between the point cloud radius corresponding to the current scan point and the point cloud radius corresponding to the next scan point. The next scan point is the next scan point on the same scan line in the same quadrant area according to the rotation direction of the lidar.

Among them, the rotation direction of the lidar can be one of counterclockwise and clockwise.

Step S304 , according to the radius difference, determine the roadside scanning points of the roadside in the target scene from the scanning points of the point cloud data.

In some embodiments of the present application, if the radius difference between the point cloud radius corresponding to the current scan point and the point cloud radius corresponding to the next scan point satisfies the difference range corresponding to the quadrant area where the current scan point is located, the current The scan point is used as the road edge scan point. By traversing the scan points on each scan line in turn, the roadside scan points in the scene can be obtained.

Specifically, assuming that the coordinate plane is the xoy plane, and the rotation direction of the lidar is counterclockwise, if it is a scanning point along the road in the first and third quadrants, then the point cloud corresponding to two adjacent scanning points in the same scan line The radius difference of the radii should satisfy r _i -r _i+1 >0; if it is a roadside scanning point in the second and fourth quadrants, then in the same scan line, the radius difference of the point cloud radii corresponding to two adjacent scanning points It should satisfy r _i -r _i+1 <0, and the radius difference calculated by the pseudo road edge (such as the arc sewer) is the opposite. Therefore, this feature can be used to filter out the pseudo road edge points such as the arc sewer, and keep the actual of the roadside scan point.

In the same way, assuming that the rotation direction of the lidar is clockwise, if it is a scanning point along the road in the first and third quadrants, then in the same scan line, the radius difference of the point cloud radii corresponding to two adjacent scanning points should satisfy r _i -r _i+1 <0; if it is a roadside scan point in the second and fourth quadrants, then in the same scan line, the radius difference of point cloud radii corresponding to two adjacent scan points should satisfy r _i -r _{i +1} > 0, while pseudo-curbs (such as curved sewers) compute the opposite for the difference in radius.

In the embodiment of the present application, the point cloud data obtained by scanning the target scene with the laser radar is obtained, and the scanning points in the point cloud data are projected to a coordinate plane in the radar coordinate system of the laser radar, so that each scan point is obtained. The projection point of the point in the coordinate plane, and then divide the coordinate plane into multiple quadrant areas, and in the same quadrant area, calculate the difference between the point cloud radii corresponding to the two adjacent scan points on the same scan line of the lidar. The radius difference between the scan points, the point cloud radius corresponding to each scan point is the length of the line segment formed by the projection point of the scan point and the coordinate origin of the coordinate plane. The radius difference has different positive and negative relationships in different quadrants. According to the radius difference, the terminal device can determine the roadside scanning points of the roadside in the target scene from the scanning points of the point cloud data, so as to avoid the false roadside points being judged as The roadside scanning point improves the accuracy of roadside point detection.

In order to reduce the interference of obvious non-roadside points such as walls and railings to roadside detection, in some embodiments of the present application, before calculating the radius difference, the terminal device may determine the ground point in the scanning points of the point cloud data.

Specifically, the terminal device can calculate the included angle between the line connecting the two scanning points on the two adjacent scanning lines of the lidar and the coordinate plane in the same rotational azimuth of the lidar, and then, according to the angle of the included angle Angle, the ground point is determined from the scan points of the point cloud data.

Among them, the rotation azimuth is the perception direction of a certain rotation angle of the lidar during the rotation process. For example, each laser beam of the lidar can perform a 360° rotation scan to obtain the scan line corresponding to each laser beam. During the rotation process, from Starting from 0°, a scan point is obtained by scanning every 5°, and the rotation orientation can refer to the sensing direction rotated by 0°, the sensing direction rotated by 5°, ..., the sensing direction rotated by 355°. Since each scan line in each sensing direction has a corresponding scan point, according to the scan sequence number of the scan point, the scan point with the same scan sequence number is also the scan point in the same rotational orientation.

Therefore, based on the scanning sequence number and line number of the scanning point, the terminal device can obtain two scanning points on the same rotational azimuth, which are located on two adjacent scanning lines of the lidar, and then calculate the same rotational azimuth of the lidar, respectively. The angle between the line connecting two scan points on two adjacent scan lines of the lidar and the coordinate plane.

It should be understood that the height difference between the scanning points of objects such as road edges and road surfaces and the coordinate plane is within a certain range, so the first included angle between the line connecting the scanning points on the road surface and/or the road edge and the coordinate plane is also within a certain range. within a certain range. For objects such as walls and railings that are significantly higher than the curb and road surface, the second included angle between the scanning point and the scanning point on the road or the curb and the coordinate plane will be significantly larger than the first included angle.

Therefore, the terminal device can obtain the angle threshold that is preset according to the empirical value. If the angle of the included angle is smaller than the angle threshold, the two scanning points corresponding to this included angle are determined as ground points. If the angle is greater than or equal to the angle threshold, at least one of the two scanning points corresponding to the included angle is determined as a non-ground point.

It should be noted that the ground point here includes the scanning points of the aforementioned road, road edge, and pseudo-road edge. Only scan points of obvious non-roadside objects such as walls and railings can be filtered out here.

At this point, the terminal device can calculate the radius difference between the point cloud radii corresponding to the two adjacent ground points on the same scan line of the lidar in the same quadrant area, avoiding obvious damage to walls, railings, etc. The scanning points of non-roadside objects are processed to reduce the amount of calculation and improve the accuracy of roadside detection.

Considering that in the process of radar scanning, the farther the scanning distance is, the more blurred the point cloud features are, and the height difference between the road edge and the road is about 10 to 15 cm. Based on this, the terminal device can also remove point clouds with inconspicuous features in the ground points.

Specifically, as shown in FIG. 4 , the terminal device removing the point cloud with inconspicuous features in the ground points may include the following steps S401 to S403 .

Step S401, obtaining a first distance between the lidar and the road in the target scene.

In some embodiments of the present application, the above-mentioned lidar can be installed on the roof of the vehicle, the installation position of the lidar is known, and the first distance corresponding to the vehicle traveling on the road is the installation height of the lidar. In other embodiments of the present application, the terminal device can also identify the ground through the identification algorithm, and calculate the first distance between the lidar and the road in the target scene.

Step S402, according to the first distance, determine the radar projection point of the lidar on the road.

It should be understood that, based on the coordinate position of the lidar in the radar coordinate system and the above-mentioned first distance, the coordinate position of the radar projection point in the radar coordinate system can be calculated.

Step S403, according to the radar projection point and the coordinate position of each ground point in the radar coordinate system, screen out the ground points whose distance from the radar projection point is greater than the distance threshold.

Specifically, the ground points whose distance from the radar projection point in each coordinate direction is greater than the distance threshold can be screened out respectively. That is, point clouds whose distances in the x, y, and z directions from the radar projection point are greater than the distance threshold are screened out. Among them, the distance threshold can be adjusted according to the actual situation.

In this way, the terminal device can screen out point clouds with a large distance from the radar and the ground, and then remove point clouds with inconspicuous features, which can reduce the amount of calculation and improve the accuracy of road edge detection.

In order to further improve the accuracy of roadside point detection, in some embodiments of the present application, as shown in FIG. 5 , the above step S304 may specifically include the following steps S501 to S503 .

Step S501, if the radius difference between the point cloud radius corresponding to the current scan point and the point cloud radius corresponding to the next scan point satisfies the difference range corresponding to the quadrant area where the current scan point is located, then the current scan point is used as the road edge scan. point as the first curb candidate point.

In some embodiments of the present application, the terminal device may determine the first roadside candidate point according to the radius difference, and then determine the roadside scan point from the first roadside candidate points.

It should be noted that, for the method of determining the first roadside candidate point, reference may be made to the description of step S304, which will not be repeated in this application.

Step S502: Calculate the slope of a straight line formed by every two adjacent first paths along the candidate points on the same scanning line of the lidar in turn.

The slope can be calculated according to the coordinate positions of every two adjacent first roadside candidate points on the same scan line.

Step S503 , according to the slope, determine the outgoing scanning point from the first candidate points of the wayside.

Specifically, if the slope difference between multiple slopes corresponding to N consecutive first roadside candidate points on the same scan line is less than the difference threshold, the consecutive N first roadside candidate points are used as roadside scanning points, wherein , N is greater than 2. By traversing the first roadside candidate points on each scan line in turn, the roadside scan points in the scene can be obtained.

It should be understood that if it is a road edge point feature, in the same scan line, the connection between the first candidate point and the n first candidate points before and after should be smooth, and the slope difference is very small, so this feature can be passed. Eliminate some noise points in the candidate points of the first road edge.

Further, as shown in FIG. 6 , in some embodiments of the present application, the above step S503 may further include the following steps S601 to S603.

Step S601 , if the slope difference between multiple slopes corresponding to N consecutive first roadside candidate points on the same scan line is smaller than the difference threshold, the consecutive N first roadside candidate points are used as the second candidate point.

In some embodiments of the present application, the terminal device may determine the second roadside candidate point according to the slope, and then determine the roadside scanning point from the second roadside candidate points.

It should be noted that, for the determination method of the second roadside candidate point, reference may be made to the description of step S503, which is not repeated in this application.

Step S602 , in each quadrant region in turn, determine the second candidate point that is closest to the laser radar on each scan line.

Specifically, the terminal device may determine the distance between the scanning point and the lidar based on the coordinates of the scanning point in the radar coordinate system.

Step S603, determining a roadside scanning point according to the determined second candidate point.

In the embodiment of the present application, the terminal device may divide the second candidate point into four quadrant areas, arrange them in sequence according to different scan lines, and search each scan line in each quadrant area for the point closest to the lidar in each scan line. as a road scan point.

Considering the situation of dynamic targets in complex scenes, as shown in FIG. 7 , the above step S603 may include the following steps S701 to S702.

Step S701, taking the determined second candidate point as the third candidate point.

In some embodiments of the present application, the terminal device may determine the third roadside candidate point according to the distance from the lidar, and then determine the roadside scanning point from the third roadside candidate points.

It should be noted that, for the determination method of the third roadside candidate point, reference may be made to the description of step S603, which is not repeated in this application.

Step S702 , perform straight line fitting on the third candidate point in each quadrant area, and use the non-outlier point in the third candidate point as the road edge scanning point according to the straight line obtained by fitting.

Specifically, the terminal device may use a random sampling consistency (Random Sample Consensus, RANSAC) straight line fitting method in each quadrant area to eliminate outliers. Other straight-line fitting methods are also applicable to this application, which will not be repeated in this application.

It should be understood that the scanning points of dynamic targets (such as the scanning points of tires) are obviously outliers compared with roadside points. Combining the above strategies and straight line fitting, outliers and dynamic points can be effectively eliminated, while retaining more radians. Large curved roadside features, get the final accurate roadside scan points.

Please refer to FIG. 8 . FIG. 8 shows a schematic diagram of a specific flow of road edge detection in the present application.

After the terminal obtains the point cloud data, it can determine the road point according to the angle between the connection line between the two scan points on the two adjacent scan lines in the same rotational orientation and the coordinate plane, and then perform four kinds of methods on the road point. The point screening strategy is used to determine candidate roadside points (third roadside points). Among them, the four road edge point screening strategies are: range-based filtering strategy, filtering strategy based on quadrant division and radius difference, filtering strategy based on slope, and filtering strategy based on quadrant division and nearest point screening. For each strategy, reference may be made to the descriptions of FIG. 4 , FIG. 1 , FIG. 5 , and FIG. 6 in sequence, which will not be repeated in this application. For the candidate roadside points, the terminal device can perform straight line fitting, remove outliers, and obtain the final roadside scan points. Reference can be made to the description of FIG. 7 here, which will not be repeated in this application. The above-mentioned road edge detection method can effectively prevent the road edge false detection problem in scenarios including arc-shaped sewers, and the road edge false detection problem in scenarios including dynamic targets such as wide intersections, which is universal in complex outdoor scenes. higher.

It should be understood that, in practical applications, the terminal device may execute each policy in turn according to the process shown in FIG. 8 , or may execute one of the policies alone, or execute any combination of these policies. No restrictions.

It should be noted that, for the sake of simple description, the foregoing method embodiments are all expressed as a series of action combinations, but those skilled in the art should know that the present application is not limited by the described action sequence. Because according to the present application, certain steps may be performed in other orders.

FIG. 9 is a schematic structural diagram of a road edge detection apparatus 900 according to an embodiment of the present application, where the road edge detection apparatus 900 is configured on a terminal device.

Specifically, the road edge detection device 900 may include:

A point cloud data acquisition unit 901, configured to acquire point cloud data obtained by scanning a target scene with a lidar;

A scanning point projection unit 902, configured to project the scanning points in the point cloud data to a coordinate plane in the radar coordinate system of the lidar, to obtain the projection point of each scanning point in the coordinate plane ;

The radius difference calculation unit 903 is configured to divide the coordinate plane into a plurality of quadrant regions, and in the same quadrant region, calculate the two adjacent scan points on the same scan line of the lidar, respectively The radius difference between the corresponding point cloud radii, the point cloud radius corresponding to each of the scanning points is the length of the line segment formed by the projection point of the scanning point and the coordinate origin of the coordinate plane;

The roadside point determination unit 904 is configured to determine, according to the radius difference, a roadside scan point of the roadside in the target scene from the scan points of the point cloud data.

In some embodiments of the present application, the above-mentioned road edge detection apparatus 900 may include a ground point detection unit, which is used for calculating the same rotational azimuth of the lidar, which are respectively located on two adjacent scan lines of the lidar. The included angle between the connecting line of the two scanning points and the coordinate plane; according to the angle of the included angle, the ground point is determined from the scanning points of the point cloud data. Correspondingly, the above-mentioned radius difference calculation unit 903 can be used to: in the same quadrant area, calculate the distance between the point cloud radii corresponding to the two adjacent ground points on the same scan line of the lidar, respectively. radius difference.

In some embodiments of the present application, the above-mentioned road edge detection apparatus 900 may include a screening unit, which is specifically configured to: obtain a first distance between the lidar and the road in the target scene; The radar projection point of the lidar on the road; according to the radar projection point and the coordinate position of each of the ground points in the radar coordinate system, screen out the distance to the radar projection point that is greater than the distance Threshold ground point.

In some embodiments of the present application, the above-mentioned radius difference calculation unit 903 may be specifically configured to: sequentially take each of the scan points as the current scan point, and calculate the point cloud radius corresponding to the current scan point and the corresponding point of the next scan point The radius difference between the radii of the point clouds, wherein the next scan point is the next scan point on the same scan line in the same quadrant area according to the rotation direction of the lidar. point.

In some embodiments of the present application, the above-mentioned roadside point determination unit 904 may be specifically configured to: if the radius difference between the point cloud radius corresponding to the current scan point and the point cloud radius corresponding to the next scan point satisfies The difference value range corresponding to the quadrant area where the current scan point is located, the current scan point is used as the roadside scan point; or, if the point cloud radius corresponding to the current scan point and the next scan point The radius difference between the corresponding point cloud radii satisfies the difference range corresponding to the quadrant area where the current scan point is located, then the current scan point is taken as the roadside scan point as the first roadside candidate point; the slope of the straight line formed by every two adjacent candidate points of the first road along the same scanning line of the lidar; according to the slope, the road edge is determined from the candidate points of the first road Scan point.

In some embodiments of the present application, the above-mentioned roadside point determination unit 904 may be specifically configured to: if the slope difference between a plurality of the slopes corresponding to the N consecutive first roadside candidate points on the same scan line is less than the difference value threshold, then the consecutive N candidate points of the first roadside are used as the roadside scanning points, where N is greater than 2; or, if N consecutive candidate points of the first roadside on the same scan line correspond to The slope difference between a plurality of the slopes is less than the difference threshold, then the consecutive N first roadside candidate points are used as the second candidate points; in each of the quadrant regions, determine each scan line in turn The second candidate point with the closest distance to the lidar on the top of the road is determined; the roadside scanning point is determined according to the determined second candidate point.

In some embodiments of the present application, the above-mentioned roadside point determination unit 904 may be specifically configured to: use the determined second candidate point as the roadside scanning point; or, use the determined second candidate point as the roadside scanning point point as the third candidate point; perform straight line fitting on the third candidate point in each quadrant area, and use the non-outlier points in the third candidate point as the Describe the scan point along the road.

It should be noted that, for the convenience and brevity of description, the specific working process of the above-mentioned road edge detection device 900 may refer to the corresponding processes of the methods described in FIGS.

As shown in FIG. 10 , it is a schematic diagram of a terminal device according to an embodiment of the present application. The terminal device 10 may include: a processor 100, a memory 101, and a computer program 102 stored in the memory 101 and executable on the processor 100, such as a road edge detection program. When the processor 100 executes the computer program 102 , the steps in each of the above embodiments of the road edge detection method are implemented, for example, steps S301 to S304 shown in FIG. 3 . Alternatively, when the processor 100 executes the computer program 102, the functions of the modules/units in the above-mentioned device embodiments are realized, for example, the point cloud data acquisition unit 901, the scanning point projection unit 902, the radius difference calculation shown in FIG. 9 unit 903 and waypoint determination unit 904.

The computer program may be divided into one or more modules/units, which are stored in the memory 101 and executed by the processor 100 to complete the present application. The one or more modules/units may be a series of computer program instruction segments capable of performing specific functions, and the instruction segments are used to describe the execution process of the computer program in the terminal device.

For example, the computer program may be divided into: a point cloud data acquisition unit, a scan point projection unit, a radius difference calculation unit, and a roadside point determination unit.

The specific functions of each unit are as follows: a point cloud data acquisition unit, which is used to acquire point cloud data obtained by scanning a target scene by a lidar; a scanning point projection unit, which is used to project the scanning points in the point cloud data to the laser A coordinate plane in the radar coordinate system of the radar, to obtain the projection point of each scanning point in the coordinate plane; the radius difference calculation unit is used to divide the coordinate plane into a plurality of quadrant areas, and in the same In the quadrant area, the radius difference between the point cloud radii corresponding to the two adjacent scan points on the same scan line of the lidar is calculated, and the point cloud radius corresponding to each scan point is: The length of the line segment formed by the projection point of the scan point and the coordinate origin of the coordinate plane; a roadside point determination unit, configured to determine the scan point from the point cloud data according to the radius difference The curb scan point of the curb within the target scene.

The terminal device may include, but is not limited to, the processor 100 and the memory 101 . Those skilled in the art can understand that FIG. 10 is only an example of a terminal device, and does not constitute a limitation to the terminal device. It may include more or less components than the one shown in the figure, or combine some components, or different components, such as The terminal device may also include an input and output device, a network access device, a bus, and the like.

The so-called processor 100 may be a central processing unit (Central Processing Unit, CPU), and may also be other general-purpose processors, digital signal processors (Digital Signal Processors, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), Off-the-shelf programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, and the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

The memory 101 may be an internal storage unit of the terminal device, such as a hard disk or a memory of the terminal device. The memory 101 may also be an external storage device of the terminal device, such as a pluggable hard disk, a smart memory card (Smart Media Card, SMC), a Secure Digital (Secure Digital, SD) card equipped on the terminal device, Flash card (Flash Card) and so on. Further, the memory 101 may also include both an internal storage unit of the terminal device and an external storage device. The memory 101 is used to store the computer program and other programs and data required by the terminal device. The memory 101 may also be used to temporarily store data that has been output or will be output.

It should be noted that, for the convenience and brevity of description, the structure of the above-mentioned terminal device may also refer to the specific description of the structure in the method embodiment, which will not be repeated here.

Those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. Module completion, that is, dividing the internal structure of the device into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiment may be integrated in one processing unit, or each unit may exist physically alone, or two or more units may be integrated in one unit, and the above-mentioned integrated units may adopt hardware. It can also be realized in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing from each other, and are not used to limit the protection scope of the present application. For the specific working processes of the units and modules in the above-mentioned system, reference may be made to the corresponding processes in the foregoing method embodiments, which will not be repeated here.

In the foregoing embodiments, the description of each embodiment has its own emphasis. For parts that are not described or described in detail in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Skilled artisans may use different methods for implementing the described functionality for each particular application, but such implementations should not be considered beyond the scope of this application.

In the embodiments provided in this application, it should be understood that the disclosed apparatus/terminal device and method may be implemented in other manners. For example, the apparatus/terminal device embodiments described above are only illustrative. For example, the division of the modules or units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units. Or components may be combined or may be integrated into another system, or some features may be omitted, or not implemented. On the other hand, the shown or discussed mutual coupling or direct coupling or communication connection may be through some interfaces, indirect coupling or communication connection of devices or units, and may be in electrical, mechanical or other forms.

The units described as separate components may or may not be physically separated, and components displayed as units may or may not be physical units, that is, may be located in one place, or may be distributed to multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution in this embodiment.

In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically alone, or two or more units may be integrated into one unit. The above-mentioned integrated units may be implemented in the form of hardware, or may be implemented in the form of software functional units.

The integrated modules/units, if implemented in the form of software functional units and sold or used as independent products, may be stored in a computer-readable storage medium. Based on this understanding, the present application can implement all or part of the processes in the methods of the above embodiments, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer When the program is executed by the processor, the steps of the foregoing method embodiments can be implemented. Wherein, the computer program includes computer program code, and the computer program code may be in the form of source code, object code, executable file or some intermediate form, and the like. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, removable hard disk, magnetic disk, optical disk, computer memory, Read-Only Memory (ROM) , Random Access Memory (Random Access Memory, RAM), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable media may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction, for example, in some jurisdictions, according to legislation and patent practice, the computer-readable media Electric carrier signals and telecommunication signals are not included.

The above-mentioned embodiments are only used to illustrate the technical solutions of the present application, but not to limit them; although the present application has been described in detail with reference to the above-mentioned embodiments, those of ordinary skill in the art should understand that: it can still be used for the above-mentioned implementations. The technical solutions described in the examples are modified, or some technical features thereof are equivalently replaced; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions in the embodiments of the application, and should be included in the within the scope of protection of this application.

Claims (10)

1. a road edge detection method, is characterized in that, comprises: Obtain point cloud data obtained by scanning the target scene with lidar; Projecting the scanning points in the point cloud data to a coordinate plane in the radar coordinate system of the lidar to obtain a projection point of each of the scanning points in the coordinate plane; Divide the coordinate plane into a plurality of quadrant areas, and in the same quadrant area, calculate the distance between the point cloud radii corresponding to the two adjacent scan points on the same scan line of the lidar. Radius difference, the point cloud radius corresponding to each of the scanning points is the length of the line segment formed by the projection point of the scanning point and the coordinate origin of the coordinate plane; According to the radius difference, a roadside scan point of a roadside in the target scene is determined from the scan points of the point cloud data. 2 . The road edge detection method according to claim 1 , wherein in the same quadrant area, two adjacent scanning points on the same scanning line of the lidar are calculated respectively. 3 . Before the radius difference between the corresponding point cloud radii, the road edge detection method includes: Calculate the angle between the line connecting the two scanning points on the two adjacent scanning lines of the lidar and the coordinate plane in the same rotational azimuth of the lidar; According to the angle of the included angle, the ground point is determined from the scanning points of the point cloud data; In the same quadrant area, calculating the radius difference between the point cloud radii corresponding to the two adjacent scan points on the same scan line of the lidar, including: In the same quadrant area, the radius difference between the point cloud radii corresponding to the two adjacent ground points on the same scan line of the lidar is calculated. 3 . The road edge detection method according to claim 2 , wherein after the ground point is determined from the scanning points of the point cloud data according to the angle of the included angle, the road edge detection method is 3. 4 . Methods also include: obtaining the first distance between the lidar and the road in the target scene; determining the radar projection point of the lidar on the road according to the first distance; According to the radar projection point and the coordinate position of each of the ground points in the radar coordinate system, the ground points whose distance from the radar projection point is greater than a distance threshold are screened out. 4. The road edge detection method according to any one of claims 1 to 3, characterized in that, in the same quadrant area, two adjacent locations on the same scan line of the lidar are calculated. The radius difference between the point cloud radii corresponding to the scanning points, including: Take each of the scanning points as the current scanning point in turn, and calculate the radius difference between the point cloud radius corresponding to the current scanning point and the point cloud radius corresponding to the next scanning point, wherein the next scanning point is according to The rotation direction of the lidar, the current scan point is the next scan point on the same scan line in the same quadrant area. 5 . The road edge detection method according to claim 4 , wherein the road edge of the road edge in the target scene is determined from the scanning points of the point cloud data according to the radius difference. 6 . Scan points, including: If the radius difference between the point cloud radius corresponding to the current scan point and the point cloud radius corresponding to the next scan point satisfies the difference range corresponding to the quadrant area where the current scan point is located, the current scan point is point as the roadside scanning point; or, If the radius difference between the point cloud radius corresponding to the current scan point and the point cloud radius corresponding to the next scan point satisfies the difference range corresponding to the quadrant area where the current scan point is located, the current scan point is point as the roadside scanning point as the first roadside candidate point; Calculate the slope of the straight line formed by every two adjacent first road along the candidate points on the same scanning line of the lidar in turn; According to the slope, the roadside scan point is determined from the first roadside candidate points. 6. The roadside detection method according to claim 5, wherein the determining the roadside scanning point from the first roadside candidate points according to the slope comprises: If the slope difference between the slopes corresponding to the N consecutive first roadside candidate points on the same scan line is smaller than the difference threshold, the consecutive N first roadside candidate points are used as the road along the scan point, where N is greater than 2; or, If the slope difference between the slopes corresponding to the N consecutive first roadside candidate points on the same scan line is smaller than the difference threshold, the consecutive N first roadside candidate points are used as the second candidate point; In each of the quadrant regions in turn, determine the second candidate point that is closest to the lidar on each scan line; The roadside scanning point is determined according to the determined second candidate point. 7. The roadside detection method according to claim 6, wherein the determining the roadside scanning point according to the determined second candidate point comprises: Using the determined second candidate point as the roadside scanning point; or, using the determined second candidate point as a third candidate point; A straight line is performed on the third candidate point in each of the quadrant regions, and according to the straight line obtained by fitting, a non-outlier point in the third candidate point is used as the roadside scanning point. 8. A road edge detection device, characterized in that, comprising: The point cloud data acquisition unit is used to acquire the point cloud data obtained by scanning the target scene with the lidar; a scanning point projection unit, configured to project the scanning points in the point cloud data to a coordinate plane in the radar coordinate system of the lidar to obtain a projection point of each of the scanning points in the coordinate plane; a radius difference calculation unit, configured to divide the coordinate plane into a plurality of quadrant areas, and in the same quadrant area, calculate the corresponding corresponding scan points of the two adjacent scan points on the same scan line of the lidar respectively The radius difference between the point cloud radii, the point cloud radius corresponding to each scan point is the length of the line segment formed by the projection point of the scan point and the coordinate origin of the coordinate plane; A roadside point determination unit, configured to determine a roadside scan point of a roadside in the target scene from the scan points of the point cloud data according to the radius difference. 9. A terminal device, comprising a memory, a processor and a computer program stored in the memory and running on the processor, wherein the processor implements the computer program as claimed in the claims when executing the computer program The steps of any one of 1 to 7 of the road edge detection method. 10. A computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, characterized in that, when the computer program is executed by a processor, the road edge detection according to any one of claims 1 to 7 is realized steps of the method.

Images