CN113917487A - Laser radar-based method for detecting road edge and drivable area of closed road - Google Patents
Laser radar-based method for detecting road edge and drivable area of closed road Download PDFInfo
- Publication number
- CN113917487A CN113917487A CN202111013260.9A CN202111013260A CN113917487A CN 113917487 A CN113917487 A CN 113917487A CN 202111013260 A CN202111013260 A CN 202111013260A CN 113917487 A CN113917487 A CN 113917487A
- Authority
- CN
- China
- Prior art keywords
- road edge
- point cloud
- cloud data
- grid
- voxel
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S17/00—Systems using the reflection or reradiation of electromagnetic waves other than radio waves, e.g. lidar systems
- G01S17/88—Lidar systems specially adapted for specific applications
- G01S17/93—Lidar systems specially adapted for specific applications for anti-collision purposes
- G01S17/931—Lidar systems specially adapted for specific applications for anti-collision purposes of land vehicles
Abstract
The invention provides a method for detecting a closed road edge and a drivable area based on a laser radar, which comprises the following steps: acquiring laser radar detection point cloud data; detecting point cloud data based on voxel grid grouping; the voxel grid is obtained by dividing a set detection space; combining grids formed by projecting voxel grids on a ground surface plane to serve as a grid map, and binarizing and marking the grid map according to the height distribution of the detection point cloud data in the voxel grids; and determining a road edge and a travelable area according to the grid map of the binary mark. According to the invention, the point cloud data is detected in groups through the voxel grids, so that the point cloud data in each voxel grid can be respectively subjected to subsequent operation in parallel, and the operation efficiency of road edge and travelable area detection is improved; meanwhile, binarization marking is carried out based on the height distribution of the point cloud data detected in the voxel grid, and a road edge and a travelable area are determined through a grid map of the binarization marking, so that the operation steps are further simplified, and the requirement of computing resources is reduced.
Description
Technical Field
The invention relates to the technical field of automatic driving, in particular to a method for detecting a closed road edge and a drivable area based on a laser radar.
Background
Road detection has been considered as a key technology in the field of automatic driving and has attracted attention of a large number of researchers. However, automated driving decisions based solely on road detection may still fail to address some emergencies. In fact, while driving a car, a human driver understands a scene by classifying obstacles from non-obstacles, not just identifying roads. For autonomous driving, detecting such "flat areas" rather than road areas may provide more comprehensive knowledge for the decision making process, making autonomous vehicle behavior closer to human drivers.
At present, the mainstream travelable region detection method is to perform semantic segmentation on a travelable region based on a two-dimensional RGB image, and a detection result of a three-dimensional travelable region needs to be obtained by converting 2D pixels into 3D point clouds through a result of an image depth estimation task and adding camera calibration parameters and performing post-processing.
The 3D drivable area is more practical for planning and controlling the automatic driving than the 2D drivable area. Although the existing method is mature in 2D semantic segmentation technology for the travelable region and high in accuracy, the travelable region can be converted into a real-world 3D coordinate system only by depending on depth estimation and camera calibration parameters, and the accumulated errors cause the obtained 3D travelable region to be inaccurate. And if a travelable area of 360 degrees is to be obtained, a plurality of groups of camera results are needed for post-fusion, post-processing is complex, and time consumption is high.
In order to solve the problems, the prior art provides methods for obtaining a road edge detection result by screening characteristics of scanning lines of mechanical rotary laser radar point clouds and clustering the point clouds.
However, with the popularization of low-cost solid-state lidar, the efficiency of the existing road edge detection algorithm based on mechanical rotation type lidar faces a great challenge. The point cloud generated by the solid laser radar is denser than that of the traditional mechanical rotary laser radar and has no characteristics of ring scanning lines, so that the existing algorithm based on the rotary scanning lines is invalid, and the computing resource required by clustering algorithm operation on the dense point cloud is high and the computing speed is low.
Therefore, the method for detecting the road edge and the travelable area, which is suitable for the dense point cloud of the solid-state laser radar, is quicker and more efficient, and has higher necessity and practical value.
Disclosure of Invention
The invention provides a method for detecting a road edge and a drivable area of a closed road based on a laser radar, which is used for overcoming the defects of high calculation resource and low calculation speed required by dense point cloud clustering in the prior art and realizing efficient and accurate detection of the edge and the drivable area.
The invention provides a method for detecting a closed road edge and a drivable area based on a laser radar, which comprises the following steps:
acquiring laser radar detection point cloud data;
grouping the detection point cloud data based on voxel grid; the voxel grid is obtained by dividing a set detection space;
combining grids formed by projection of the voxel grids on the ground surface plane to serve as grid images, and carrying out binarization marking on the grid images according to the height distribution of detection point cloud data in the voxel grids;
and determining a road edge and a travelable area according to the grid map of the binary mark.
According to the method for detecting the road edge and the travelable area of the closed road based on the laser radar, the step of determining the road edge and the travelable area according to the grid map marked by the binaryzation mark comprises the following steps of:
determining a road edge endpoint according to the endpoint region set in the grid map;
moving a road edge sliding window established in the grid map from the road edge end point, and determining a road edge according to a binarization marking value of a grid in the road edge sliding window;
and determining a travelable area according to the road edge.
According to the method for detecting the road edge and the travelable area of the closed road based on the laser radar, the step of determining the road edge end point according to the end point area set in the grid map comprises the following steps:
establishing a two-dimensional Cartesian coordinate system on the grid map by taking a current vehicle as an origin and taking the advancing direction of the current vehicle as a y-axis, and establishing an endpoint function aiming at a set part of the grid map; the value of the endpoint function is the sum of the binaryzation mark values of the grids with the same x coordinate;
determining the first position and the second position as road edge endpoints according to the endpoint function;
the first position has coordinates of (x)1，0)，x1The minimum value in the x coordinate set corresponding to the extreme value of the endpoint function in the x-axis positive half shaft is obtained;
the second position has coordinates of (x)2，0)，x2The maximum value in the x coordinate set corresponding to the extreme value of the endpoint function in the negative half axis of the x axis.
According to the method for detecting the road edge and the travelable area of the closed road based on the laser radar, the step of moving the road edge sliding window established in the grid map from the end point of the road edge and determining the road edge according to the binary marking value of the grid in the road edge sliding window comprises the following steps:
establishing a rectangular road edge sliding window in the grid graph, and setting the position of the road edge sliding window as the position where the x coordinate of the center point of the road edge sliding window is the same as the x coordinate of the end point of the road edge;
updating the position of the road edge sliding window according to the binarization label value of each grid in the road edge sliding window, and adding the center point of the road edge sliding window after the position is updated to a road edge point set;
determining the extension direction of the road edge according to the binaryzation mark value of each grid in the road edge sliding window after the position is updated, and moving the road edge sliding window according to the extension direction of the road edge;
returning to the step of updating the position of the road edge sliding window according to the binarization label value of each grid in the road edge sliding window and adding the center point of the road edge sliding window after the position is updated to the road edge point set until a set stop condition is met;
and determining the road edge according to the road edge point set fitting.
According to the method for detecting the closed road edge and the travelable area based on the laser radar, the step of determining the travelable area according to the road edge comprises the following steps:
dividing the grid map into three parts by taking the two road edges as boundaries, and marking the part where the current vehicle is located as a road area;
determining a part excluding an obstacle area in the road area as a travelable area;
the barrier area is a part of the first area excluding the second area; the first area is formed by encircling an obstacle tangent line and a road edge in the raster image; the second area is formed by encircling an obstacle tangent and an obstacle outline in the grid map;
the obstacle tangent line is a tangent line of an obstacle curve passing through the position of the current vehicle; the obstacle curve is one or more curves formed by combining obstacle projection points; the obstacle projection point is a part of the projection point of the detection point cloud data on the ground surface plane, which is positioned in the road area;
the obstacle contour is a part close to the position of the current vehicle in two parts formed by dividing the obstacle curve by two obstacle tangent points; the obstacle tangent point is a tangent point of the obstacle tangent line and the obstacle curve.
According to the method for detecting the road edge and the travelable area of the closed road based on the laser radar, the step of binaryzation marking of the grid map according to the height distribution of the point cloud data detected in the voxel grid comprises the following steps:
calculating the height variance and the three-dimensional surface curvature of the detection point cloud data in the voxel grid;
if the height variance of the detection point cloud data in the voxel grid is greater than a set variance threshold value and the three-dimensional surface curvature of the detection point cloud data in the voxel grid is greater than a set curvature threshold value, marking a grid formed by projection of the voxel grid on a ground surface plane as 1;
and if the height variance of the detection point cloud data in the voxel grid is not greater than a set variance threshold value, or the three-dimensional surface curvature of the detection point cloud data in the voxel grid is not greater than a set curvature threshold value, marking a grid formed by projection of the voxel grid on a ground surface plane as 0.
According to the method for detecting the road edge and the travelable area of the closed road based on the laser radar, the three-dimensional surface curvature sigma meets the following requirements:
in the formula, λ0、λ1、λ2Is the three eigenvectors of the covariance matrix C, and0<λ1<λ2(ii) a The covariance matrix C satisfies:
in the formula, k is the number of the detection point cloud data in the voxel grid; pi is the coordinate of the ith detection point cloud data in the voxel grid;
and the mean value of the coordinate of the detected point cloud data in the voxel grid is obtained.
According to the method for detecting the road edge and the travelable area of the closed road based on the laser radar, the step of acquiring the detection point cloud data of the laser radar comprises the following steps:
acquiring original point cloud data of a laser radar;
selecting single-frame point cloud data from the original point cloud data to eliminate ground point cloud and form single-frame detection point cloud data;
and superposing the single-frame detection point cloud data of a set number of adjacent frames to form detection point cloud data.
According to the method for detecting the road edge and the drivable area of the closed road based on the laser radar, the step of selecting single-frame point cloud data from the original point cloud data to eliminate ground point cloud and then forming the single-frame detection point cloud data comprises the following steps:
determining a ground surface plane; the ground surface plane is a verification band with the maximum judgment weight; the verification belt is formed by translating a verification plane along the normal direction and is provided with a verification area with a set thickness; the verification plane is obtained by randomly sampling a set number of point cloud fits in single-frame point cloud data; the judgment weight is the proportion of the number of point clouds in the verification band to the number of point clouds outside the verification band;
and eliminating the point cloud data positioned in the ground surface plane from the single-frame point cloud data to form single-frame detection point cloud data.
According to the method for detecting the closed road edge and the travelable area based on the laser radar, the step of determining the road edge comprises the following steps:
correcting the first path curve and the second path curve according to the curvature of the first path curve and the curvature mean value of the second path curve;
and determining the road edge through the corrected first road edge curve and the second road edge curve.
The invention also provides a system for detecting the road edge and the drivable area of the closed road based on the laser radar, which comprises the following components:
the data acquisition module is used for acquiring laser radar detection point cloud data;
a voxel grouping module for grouping the detection point cloud data based on voxel grids; the voxel grid is obtained by dividing a set detection space;
the binarization module is used for combining grids formed by projection of the voxel grids on the ground surface plane as a grid map, and binarizing and marking the grid map according to the height distribution of the detection point cloud data in the voxel grids;
and the area determining module is used for determining a road edge and a travelable area according to the grid map of the binarization mark.
The invention also provides an electronic device, which comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the processor executes the program to realize the steps of the method for detecting the closed road edge and the travelable area based on the laser radar.
The present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of the lidar based closed road edge and travelable area detection method as described in any of the above.
The present invention also provides a computer program product comprising a computer program which, when executed by a processor, carries out the steps of the lidar-based closed road edge and travelable area detection method as defined in any of the above.
According to the method for detecting the road edge and the drivable area of the closed road based on the laser radar, the point cloud data are detected in groups through the voxel grids, so that the point cloud data in each voxel grid can be subjected to subsequent operation in parallel, and the operation efficiency of the detection of the road edge and the drivable area is improved; meanwhile, binarization marking is carried out based on the height distribution of the point cloud data detected in the voxel grid, and a road edge and a travelable area are determined through a grid map of the binarization marking, so that the operation steps are further simplified, and the requirement of computing resources is reduced.
Drawings
In order to more clearly illustrate the technical solutions of the present invention or the prior art, the drawings needed for the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
FIG. 1 is a schematic flow chart of a method for detecting a closed road edge and a travelable area based on a laser radar provided by the invention;
FIG. 2 is a schematic diagram of a first binarization grid according to an embodiment of the present invention;
fig. 3 is a schematic diagram of a second binarization grid provided by the embodiment of the invention;
FIG. 4 is a schematic diagram of signals used to determine a first position and a second position according to an embodiment of the present invention;
FIG. 5 is a schematic view of a road edge sliding window provided by an embodiment of the invention;
FIG. 6 is a schematic diagram of a first fitted path according to an embodiment of the present invention;
FIG. 7 is a second fitted edge schematic provided by an embodiment of the present invention;
FIG. 8 is a schematic view of a drivable region provided by an embodiment of the present invention;
fig. 9 is a schematic structural diagram of an electronic device provided by the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is obvious that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The method for detecting the closed road edge and the travelable area based on the laser radar of the invention is described in the following with reference to fig. 1 to 8.
As shown in fig. 1, an embodiment of the present invention provides a method for detecting a road edge and a travelable area of a closed road based on a laser radar, including:
103, grouping the detection point cloud data based on voxel grids; the voxel grid is obtained by dividing a set detection space;
105, combining grids formed by projection of the voxel grids on the ground surface plane as a grid map, and binarizing and marking the grid map according to the height distribution of the detection point cloud data in the voxel grids;
and step 107, determining a road edge and a travelable area according to the grid map of the binary mark.
The execution subject of the present embodiment may be a computer program or a computer system.
In this embodiment, the voxel grid is specifically set as follows:
first, a detection space is defined, that is, a cube with a height set on the ground surface plane with a set length and width is used as a bottom. Considering that in the field related to actual automatic driving, motor vehicles and common obstacles are concerned in the height direction, the set height reference value of the detection space is 2 meters; considering the data acquisition characteristics of the current vehicle (and the laser radar arranged on the current vehicle), one of the two sides of the length and the width of the detection space should pass through the position of the current vehicle. The specific setting of the length and width depends on the parameters of the lidar and the road conditions.
Subsequently, the detection space is divided, the subsequent steps of this embodiment rely on binarization labeling of the point cloud data in the voxel grid, and a grid map is formed by using the projection of the voxel grid on the ground surface plane, so that the height of the voxel grid can be simply set to be the same cube as the detection space, so as to avoid overlapping of the projections of the voxel grids on the ground surface plane. Further, considering the length and width of the bottom surface of the voxel grid, the present embodiment sets the length and width to 0.3m in consideration of the overall calculation accuracy (the "resolution" of the binarization grid) and the calculation efficiency.
It should be noted that the above voxel grid setting is not to be construed as a limitation to the embodiment, and this setting scheme is only a preferred scheme of the embodiment.
In some practical application scenarios, the number of voxel grids formed by the above scheme is 128 × 512, that is, point cloud data in the range of 512 × 0.3 ═ 153.6m in the forward direction of the current vehicle is considered, and point cloud data in the range of 128 × 0.3 ═ 38.4m in the left and right lateral directions of the current vehicle is considered.
The grid map in this embodiment is composed of 128 × 512 square grids according to the above arrangement of voxel grid.
The binarization label of the grid is determined by the height distribution of the point cloud data in the corresponding voxel grid, and the more discrete the height distribution of the point cloud is, the higher the probability that the voxel grid and the corresponding grid are possible to be components of the road edge is.
It should be noted that 0 and 1 in the binary flag are not to be understood as a limitation to this embodiment, and the labeling is to divide the grids into two types according to the height distribution dispersion, so that a grid higher than the threshold value may be set to 0, and a grid not higher than the threshold value may be set to 1; or two marks of 0 and 1 are not used, but other numbers, letters, colors or shades are used, and the embodiment does not limit the invention.
After the road edge is determined, the space inside the road edge can be understood as a road area, and then other vehicles or obstacles in the road can be analyzed according to the point cloud data in the road area, so that the conclusion of the travelable area is obtained.
The beneficial effect of this embodiment lies in:
according to the method for detecting the road edge and the travelable area of the closed road based on the laser radar, the point cloud data are detected in groups through the voxel grids, so that the point cloud data in each voxel grid can be subjected to subsequent operation in parallel, and the operation efficiency of the detection of the road edge and the travelable area is improved; meanwhile, binarization marking is carried out based on the height distribution of the point cloud data detected in the voxel grid, and a road edge and a travelable area are determined through a grid map of the binarization marking, so that the operation steps are further simplified, and the requirement of computing resources is reduced.
According to the above embodiment, in the present embodiment:
the step of determining the road edge and the travelable area according to the grid map of the binarization mark comprises the following steps:
determining a road edge endpoint according to the endpoint region set in the grid map;
moving a road edge sliding window established in the grid map from the road edge end point, and determining a road edge according to a binarization marking value of a grid in the road edge sliding window;
and determining a travelable area according to the road edge.
In this embodiment, the way of the curbstone sliding window is adopted, and the curbstone sliding window is traversed and moved from the starting point of the curbstone, so as to better fit the curbstone curve.
Wherein, the step of determining the road edge end point according to the end point region set in the grid map comprises the following steps:
establishing a two-dimensional Cartesian coordinate system on the grid map by taking a current vehicle as an origin and taking the advancing direction of the current vehicle as a y-axis, and establishing an endpoint function aiming at a set part of the grid map; the value of the endpoint function is the sum of the binaryzation mark values of the grids with the same x coordinate;
determining the first position and the second position as road edge endpoints according to the endpoint function;
the first position has coordinates of (x)1，0)，x1The minimum value in the x coordinate set corresponding to the extreme value of the endpoint function in the x-axis positive half shaft is obtained;
the second position has coordinates of (x)2，0)，x2The maximum value in the x coordinate set corresponding to the extreme value of the endpoint function in the negative half axis of the x axis.
In this embodiment, the cartesian coordinate system is established for convenience of explanation, and in actual use, the setting manner of the origin and the coordinate axis may be replaced, and the transformed coordinates and the coordinates in this embodiment may be obtained through simple mathematical changes.
The actual physical meaning of the road edge end point in this embodiment may be understood as a road edge starting point on two sides of the same y coordinate, which is closer to the current vehicle in the grid map; in turn, it can be understood (in some specific scenarios) that two road edge end points are farther away from the current vehicle in the grid map; it is also understood to include both the two curb starting points and the two curb ending points.
Correspondingly, in the subsequent step of traversing the road edge sliding window in this embodiment, the road edge sliding window may move from the road edge starting point, may move from the road edge end point, and may also move from the road edge starting point and the road edge end point respectively.
In this embodiment, the step of moving the road edge sliding window established in the grid map from the road edge endpoint and determining the road edge according to the binarized labeled value of the grid in the road edge sliding window includes:
establishing a rectangular road edge sliding window in the grid graph, and setting the position of the road edge sliding window as the position where the x coordinate of the center point of the road edge sliding window is the same as the x coordinate of the end point of the road edge;
updating the position of the road edge sliding window according to the binarization label value of each grid in the road edge sliding window, and adding the center point of the road edge sliding window after the position is updated to a road edge point set;
determining the extension direction of the road edge according to the binaryzation mark value of each grid in the road edge sliding window after the position is updated, and moving the road edge sliding window according to the extension direction of the road edge;
returning to the step of updating the position of the road edge sliding window according to the binarization label value of each grid in the road edge sliding window and adding the center point of the road edge sliding window after the position is updated to the road edge point set until a set stop condition is met;
and determining the road edge according to the road edge point set fitting.
In this embodiment, the size of the road edge sliding window should be larger than that of the grid, and in this embodiment, a rectangular sliding window capable of accommodating 3 × 4 — 12 grids is selected.
In the step of updating the position of the road edge sliding window according to the binarized labeled value of each grid in the road edge sliding window, the updating method may be:
and calculating the gravity center of a 1 grid (namely a grid with height discrete distribution higher than a threshold value) set in the road edge sliding window, and updating the position of the road edge sliding window to enable the center point of the road edge sliding window to be coincident with the gravity center.
In the step of determining the road edge extending direction according to the updated binarized labeled value of each grid in the road edge sliding window, the method for determining the road edge extending direction may be:
and fitting to obtain a road edge extension straight line according to a 1 grid (namely a grid with height discrete distribution higher than a threshold value) in the road edge sliding window, and determining a road edge extension direction (collinear with the road edge extension straight line) based on the road edge extension straight line according to any one or any combination of the current vehicle orientation, the starting position of the road edge sliding window and the positions of the road edge starting point/road edge terminal point.
In this embodiment, the step of determining the travelable area according to the road edge includes:
dividing the grid map into three parts by taking the two road edges as boundaries, and marking the part where the current vehicle is located as a road area;
determining a part excluding an obstacle area in the road area as a travelable area;
the barrier area is a part of the first area excluding the second area; the first area is formed by encircling an obstacle tangent line and a road edge in the raster image; the second area is formed by encircling an obstacle tangent and an obstacle outline in the grid map;
the obstacle tangent line is a tangent line of an obstacle curve passing through the position of the current vehicle; the obstacle curve is one or more curves formed by combining obstacle projection points; the obstacle projection point is a part of the projection point of the detection point cloud data on the ground surface plane, which is positioned in the road area;
the obstacle contour is a part close to the position of the current vehicle in two parts formed by dividing the obstacle curve by two obstacle tangent points; the obstacle tangent point is a tangent point of the obstacle tangent line and the obstacle curve.
The determination step of the travelable region in this embodiment can be simply understood as:
firstly, determining a road area through a road edge;
secondly, determining obstacles (possibly other running vehicles) in the road according to the three-dimensional point cloud data in the road area;
thirdly, determining the area of the blocked part of the road cataract based on Raycasting algorithm;
fourthly, in the road area, excluding the area where the road blocks the blocked part to obtain the final drivable area.
In this embodiment, the step of determining the road edge includes:
correcting the first path curve and the second path curve according to the curvature of the first path curve and the curvature mean value of the second path curve;
and determining the road edge through the corrected first road edge curve and the second road edge curve.
The beneficial effect of this embodiment lies in:
the embodiment can directly obtain the road edge and the travelable area under the accurate top view, and is simultaneously suitable for mechanical rotary and solid-state laser radars. Because the method is based on the voxelization processing, various geometric characteristics of point cloud in the voxelization are utilized, and a windowing method is used for fitting the road edge curve, the method is more efficient and has higher accuracy compared with the existing method. And then, removing the area shielded by the object under the top view through Raycasting, so that the real drivable area can be accurately acquired.
According to any of the embodiments described above, in this embodiment:
the step of binarizing and marking the grid map according to the height distribution of the detection point cloud data in the voxel grid square comprises the following steps:
calculating the height variance and the three-dimensional surface curvature of the detection point cloud data in the voxel grid;
if the height variance of the detection point cloud data in the voxel grid is greater than a set variance threshold value and the three-dimensional surface curvature of the detection point cloud data in the voxel grid is greater than a set curvature threshold value, marking a grid formed by projection of the voxel grid on a ground surface plane as 1;
and if the height variance of the detection point cloud data in the voxel grid is not greater than a set variance threshold value, or the three-dimensional surface curvature of the detection point cloud data in the voxel grid is not greater than a set curvature threshold value, marking a grid formed by projection of the voxel grid on a ground surface plane as 0.
The three-dimensional surface curvature σ satisfies:
in the formula, λ0、λ1、λ2Is the three eigenvectors of the covariance matrix C, and0<λ1<λ2(ii) a The covariance matrix C satisfies:
in the formula, k is the number of the detection point cloud data in the voxel grid; p is a radical ofiCoordinates of the ith detection point cloud data in the voxel grid;
and the mean value of the coordinate of the detected point cloud data in the voxel grid is obtained.
The beneficial effect of this embodiment lies in:
the embodiment provides a method for quantizing the distribution of three-dimensional point clouds in a voxel grid, and verification based on the curvature of a three-dimensional surface is added on the basis of height distribution, so that the calculation result is more accurate.
According to any of the embodiments described above, in this embodiment:
the step of acquiring the laser radar detection point cloud data comprises the following steps:
acquiring original point cloud data of a laser radar;
selecting single-frame point cloud data from the original point cloud data to eliminate ground point cloud and form single-frame detection point cloud data;
and superposing the single-frame detection point cloud data of a set number of adjacent frames to form detection point cloud data.
The method comprises the following steps of selecting single-frame point cloud data from the original point cloud data, removing ground point clouds, and forming single-frame detection point cloud data:
determining a ground surface plane; the ground surface plane is a verification band with the maximum judgment weight; the verification belt is formed by translating a verification plane along the normal direction and is provided with a verification area with a set thickness; the verification plane is obtained by randomly sampling a set number of point cloud fits in single-frame point cloud data; the judgment weight is the proportion of the number of point clouds in the verification band to the number of point clouds outside the verification band;
and eliminating the point cloud data positioned in the ground surface plane from the single-frame point cloud data to form single-frame detection point cloud data.
The beneficial effect of this embodiment lies in:
the method is based on the RANSAC method to fit the ground plane equation, and the point cloud is subjected to ground removing treatment, so that the interference of the ground point cloud on subsequent calculation is reduced, and the operation efficiency of the travelable area detection is effectively improved;
meanwhile, the method based on voxel grid segmentation reduces the requirement of computing resources, and therefore can support the computation of denser effective point clouds (namely, detection point cloud data after ground point clouds are eliminated), and detection point cloud data is formed by superposing single-frame detection point cloud data of a set number of adjacent frames for the calculation of subsequent steps, so that more accurate detection results of road edges and travelable areas can be obtained.
A complete embodiment, which is described in combination with actual operations as a whole, is provided below, and the present embodiment employs a lidar sensor to directly detect a travelable region under a 3D coordinate system BEV. The input of the method of the embodiment is the point cloud of the current frame and the target detection frame, and the detection frame can be obtained from a manual marking or target detection model.
The embodiment specifically comprises the following steps:
1. point cloud pretreatment: firstly, fitting a ground plane equation by using an RANSAC method, removing the ground from the point cloud, and then removing points in a point cloud target detection frame by using the output of a target detector or a manually marked detection frame. And the rest points are used as the point cloud after the point cloud of the current frame is preprocessed, and at the moment, map positioning information can be utilized to overlap the point cloud after the preprocessing of the previous 1-5 frames to the current frame, so that the dense edge characteristics of the point cloud are more obvious.
2. Detecting a road boundary line:
2.1, carrying out Pillar-column-shaped voxelization operation on the point cloud processed in the first step, such as 0.3m x0.3m of a grid, so as to obtain a map of the 3D point cloud to the 2D BEV, wherein the initialization travelable area is 1.
2.2 calculate the height variance and 3D surface curvature on the vertical ground coordinates for points within the same voxel as the geometric features of the voxel. AND then respectively adding thresholds AND performing AND operation to obtain voxels possibly having road boundaries in the form of a 2-value map under BEV.
The formula of curvature operation of the 3D surface:
Lambda(λ0、λ1、λ2) Is an eigenvector of the covariance matrix C
The curvature of the 3D surface corresponds to the weight of the smallest feature vector, the value varies from 0-1, the lower the surface is, the flatter this feature is, in general, more robust and stable than the surface normal feature.
Height variance
The height variance indicates the smoothness of the surface, and the larger the value, the more uneven the surface.
Fig. 2 and fig. 3 respectively provide grid maps after binarization marking in two different scenes, wherein a black part is a grid with a 1 and a white part is a grid with a 0.
And 2.3, accumulating the lower half part of the 2-value graph downwards to draw a signal graph, and finding out the first crest appearing at the left side and the right side of the bicycle as the starting point of the road edge.
Fig. 4 shows a signal diagram for determining a first position and a second position in an application scenario.
And 2.4 starting from the starting point, making a sliding window upwards on the 2-value graph, updating the center point of the sliding window as the gravity center of a non-0 voxel of the current window, and finally fitting a road boundary curve by using the center point of each window to obtain left and right road edge curves. Because the two side edges of the expressway scene are generally parallel to each other, the curvature can be averaged to keep the curvature of the curves of the left edge and the right edge consistent.
Fig. 5 is a schematic view of the curbstone, wherein the black rectangular blocks are the curbstone.
Fig. 6 and 7 respectively provide schematic diagrams of fitted road edges in two different scenarios.
And 2.5, setting the pixels outside the road boundary in the BEV map as 0 to represent the non-travelable area.
Raycasting, namely performing Raycasting on corner points of a target frame on a BEV image from a vehicle center point, setting an area with the sight line shielded by the target frame as 0, and setting the remaining area with the value of 1 as a drivable area.
Fig. 8 shows a schematic view of the travelable region after the racing.
The beneficial effect of this embodiment lies in:
the embodiment can directly obtain the road edge and the travelable area under the accurate top view, and is simultaneously suitable for mechanical rotary and solid-state laser radars. Because the method is based on the voxelization processing, various geometric characteristics of point cloud in the voxelization are utilized, and a windowing method is used for fitting the road edge curve, the method is more efficient and has higher accuracy compared with the existing method. And then, removing the area shielded by the object under the top view through Raycasting, so that the real drivable area can be accurately acquired. The road edges and travelable areas thus obtained can be used as the true value for neural network learning or output to a planning control module downstream of the autopilot perception.
The following describes the laser radar-based closed road edge and travelable area detection device provided by the present invention, and the laser radar-based closed road edge and travelable area detection device described below and the laser radar-based closed road edge and travelable area detection method described above may be referred to correspondingly.
The embodiment of the invention provides a laser radar-based closed road edge and travelable area detection system, which comprises:
the data acquisition module is used for acquiring laser radar detection point cloud data;
a voxel grouping module for grouping the detection point cloud data based on voxel grids; the voxel grid is obtained by dividing a set detection space;
the binarization module is used for combining grids formed by projection of the voxel grids on the ground surface plane as a grid map, and binarizing and marking the grid map according to the height distribution of the detection point cloud data in the voxel grids;
and the area determining module is used for determining a road edge and a travelable area according to the grid map of the binarization mark.
The data acquisition module comprises:
the original data submodule is used for acquiring original point cloud data of the laser radar;
the ground point cloud elimination submodule is used for selecting single-frame point cloud data from the original point cloud data to eliminate the ground point cloud so as to form single-frame detection point cloud data;
and the single-frame overlapping submodule is used for overlapping the single-frame detection point cloud data of a set number of adjacent frames to form detection point cloud data.
The ground point cloud exclusion sub-module comprises:
the earth surface plane determining unit is used for determining an earth surface plane; the ground surface plane is a verification band with the maximum judgment weight; the verification belt is formed by translating a verification plane along the normal direction and is provided with a verification area with a set thickness; the verification plane is obtained by randomly sampling a set number of point cloud fits in single-frame point cloud data; the judgment weight is the proportion of the number of point clouds in the verification band to the number of point clouds outside the verification band;
and the point cloud elimination unit is used for eliminating the point cloud data positioned in the ground surface plane from the single-frame point cloud data to form single-frame detection point cloud data.
The binarization module comprises:
the variance and curvature calculation submodule is used for calculating the height variance and the three-dimensional surface curvature of the detection point cloud data in the voxel grid;
a binarization judgment submodule for:
if the height variance of the detection point cloud data in the voxel grid is greater than a set variance threshold value and the three-dimensional surface curvature of the detection point cloud data in the voxel grid is greater than a set curvature threshold value, marking a grid formed by projection of the voxel grid on a ground surface plane as 1;
and if the height variance of the detection point cloud data in the voxel grid is not greater than a set variance threshold value, or the three-dimensional surface curvature of the detection point cloud data in the voxel grid is not greater than a set curvature threshold value, marking a grid formed by projection of the voxel grid on a ground surface plane as 0.
The three-dimensional surface curvature σ satisfies:
in the formula, λ0、λ1、λ2Is the three eigenvectors of the covariance matrix C, and0<λ1<λ2(ii) a The covariance matrix C satisfies:
in the formula, k is the number of the detection point cloud data in the voxel grid; p is a radical ofiCoordinates of the ith detection point cloud data in the voxel grid;
and the mean value of the coordinate of the detected point cloud data in the voxel grid is obtained.
The region determination module includes:
the road edge endpoint submodule is used for determining a road edge endpoint according to an endpoint region set in the grid graph;
the road edge determining submodule is used for moving a road edge sliding window established in the grid map from the road edge end point and determining a road edge according to a binary marking value of a grid in the road edge sliding window;
and the travelable area determining submodule is used for determining a travelable area according to the road edge.
The curb end submodule includes:
the system comprises an endpoint function unit, a grid map generation unit and a control unit, wherein the endpoint function unit is used for establishing a two-dimensional Cartesian coordinate system on the grid map by taking a current vehicle as an origin and taking the advancing direction of the current vehicle as a y-axis, and establishing an endpoint function aiming at a set part of the grid map; the value of the endpoint function is the sum of the binaryzation mark values of the grids with the same x coordinate;
the end point determining unit is used for determining the first position and the second position as road edge end points according to the end point function;
the first position has coordinates of (x)1，0)，x1The minimum value in the x coordinate set corresponding to the extreme value of the endpoint function in the x-axis positive half shaft is obtained;
the second position has coordinates of (x)2，0)，x2The maximum value in the x coordinate set corresponding to the extreme value of the endpoint function in the negative half axis of the x axis.
The road edge determination submodule includes:
the initial position unit of the road edge sliding window is used for establishing a rectangular road edge sliding window in the grid map and setting the position of the road edge sliding window as the position where the x coordinate of the center point of the road edge sliding window is the same as the x coordinate of the end point of the road edge;
the position updating unit of the road edge sliding window is used for updating the position of the road edge sliding window according to the binarization label value of each grid in the road edge sliding window and adding the center point of the road edge sliding window after the position is updated to a road edge point set;
the road edge sliding window moving unit is used for determining the road edge extending direction according to the updated binary marking value of each grid in the road edge sliding window and moving the road edge sliding window according to the road edge extending direction;
the sliding window iteration unit is used for returning to the step of updating the position of the road edge sliding window according to the binarization label value of each grid in the road edge sliding window and adding the center point of the road edge sliding window after the position is updated to the road edge point set until a set stop condition is met;
and the road edge fitting determination unit is used for determining the road edge according to the road edge point set fitting.
The travelable region determination submodule includes:
the road area unit is used for dividing the grid map into three parts by taking the two road edges as boundaries, and marking the part where the current vehicle is located as a road area;
an obstacle exclusion area unit configured to determine a portion excluding an obstacle area in the road area as a travelable area;
the barrier area is a part of the first area excluding the second area; the first area is formed by encircling an obstacle tangent line and a road edge in the raster image; the second area is formed by encircling an obstacle tangent and an obstacle outline in the grid map;
the obstacle tangent line is a tangent line of an obstacle curve passing through the position of the current vehicle; the obstacle curve is one or more curves formed by combining obstacle projection points; the obstacle projection point is a part of the projection point of the detection point cloud data on the ground surface plane, which is positioned in the road area;
the obstacle contour is a part close to the position of the current vehicle in two parts formed by dividing the obstacle curve by two obstacle tangent points; the obstacle tangent point is a tangent point of the obstacle tangent line and the obstacle curve.
Fig. 9 illustrates a physical structure diagram of an electronic device, and as shown in fig. 9, the electronic device may include: a processor (processor)910, a communication Interface (Communications Interface)920, a memory (memory)930, and a communication bus 940, wherein the processor 910, the communication Interface 920, and the memory 930 communicate with each other via the communication bus 940. Processor 910 may invoke logic instructions in memory 930 to perform a lidar-based enclosed road edge and drivable region detection method comprising: acquiring laser radar detection point cloud data; grouping the detection point cloud data based on voxel grid; the voxel grid is obtained by dividing a set detection space; combining grids formed by projection of the voxel grids on the ground surface plane to serve as grid images, and carrying out binarization marking on the grid images according to the height distribution of detection point cloud data in the voxel grids; and determining a road edge and a travelable area according to the grid map of the binary mark.
In addition, the logic instructions in the memory 830 may be implemented in software functional units and stored in a computer readable storage medium when the logic instructions are sold or used as independent products. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
In another aspect, the present invention also provides a computer program product, the computer program product comprising a computer program, the computer program being storable on a non-transitory computer-readable storage medium, the computer program, when executed by a processor, being capable of executing the lidar-based closed road edge and travelable area detection method provided by the above methods, the method comprising: acquiring laser radar detection point cloud data; grouping the detection point cloud data based on voxel grid; the voxel grid is obtained by dividing a set detection space; combining grids formed by projection of the voxel grids on the ground surface plane to serve as grid images, and carrying out binarization marking on the grid images according to the height distribution of detection point cloud data in the voxel grids; and determining a road edge and a travelable area according to the grid map of the binary mark.
In yet another aspect, the present invention also provides a non-transitory computer-readable storage medium having stored thereon a computer program, which when executed by a processor, implements a lidar-based closed road edge and travelable region detection method provided by the above-described methods, the method comprising: acquiring laser radar detection point cloud data; grouping the detection point cloud data based on voxel grid; the voxel grid is obtained by dividing a set detection space; combining grids formed by projection of the voxel grids on the ground surface plane to serve as grid images, and carrying out binarization marking on the grid images according to the height distribution of detection point cloud data in the voxel grids; and determining a road edge and a travelable area according to the grid map of the binary mark.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (14)
1. A method for detecting a closed road edge and a travelable area based on a laser radar is characterized by comprising the following steps:
acquiring laser radar detection point cloud data;
grouping the detection point cloud data based on voxel grid; the voxel grid is obtained by dividing a set detection space;
combining grids formed by projection of the voxel grids on the ground surface plane to serve as grid images, and carrying out binarization marking on the grid images according to the height distribution of detection point cloud data in the voxel grids;
and determining a road edge and a travelable area according to the grid map of the binary mark.
2. The lidar-based closed road edge and travelable region detection method according to claim 1, wherein the step of determining the edge and travelable region based on the grid map of binarized marks comprises:
determining a road edge endpoint according to the endpoint region set in the grid map;
moving a road edge sliding window established in the grid map from the road edge end point, and determining a road edge according to a binarization marking value of a grid in the road edge sliding window;
and determining a travelable area according to the road edge.
3. The lidar-based closed road edge and travelable area detection method according to claim 2, wherein the step of determining the edge end point according to the end point area set in the grid map comprises:
establishing a two-dimensional Cartesian coordinate system on the grid map by taking a current vehicle as an origin and taking the advancing direction of the current vehicle as a y-axis, and establishing an endpoint function aiming at a set part of the grid map; the value of the endpoint function is the sum of the binaryzation mark values of the grids with the same x coordinate;
determining the first position and the second position as road edge endpoints according to the endpoint function;
the first position has coordinates of (x)1，0)，x1The minimum value in the x coordinate set corresponding to the extreme value of the endpoint function in the x-axis positive half shaft is obtained;
the second position has coordinates of (x)2，0)，x2The maximum value in the x coordinate set corresponding to the extreme value of the endpoint function in the negative half axis of the x axis.
4. The lidar-based closed road edge and travelable region detection method according to claim 3, wherein the step of moving an edge sliding window established in the grid map from the edge end point and determining the edge according to the binarized mark value of the grid in the edge sliding window comprises:
establishing a rectangular road edge sliding window in the grid graph, and setting the position of the road edge sliding window as the position where the x coordinate of the center point of the road edge sliding window is the same as the x coordinate of the end point of the road edge;
updating the position of the road edge sliding window according to the binarization label value of each grid in the road edge sliding window, and adding the center point of the road edge sliding window after the position is updated to a road edge point set;
determining the extension direction of the road edge according to the binaryzation mark value of each grid in the road edge sliding window after the position is updated, and moving the road edge sliding window according to the extension direction of the road edge;
returning to the step of updating the position of the road edge sliding window according to the binarization label value of each grid in the road edge sliding window and adding the center point of the road edge sliding window after the position is updated to the road edge point set until a set stop condition is met;
and determining the road edge according to the road edge point set fitting.
5. The lidar-based closed road curb and drivable area detection method of claim 2, wherein the step of determining a drivable area based on the curb comprises:
dividing the grid map into three parts by taking the two road edges as boundaries, and marking the part where the current vehicle is located as a road area;
determining a part excluding an obstacle area in the road area as a travelable area;
the barrier area is a part of the first area excluding the second area; the first area is formed by encircling an obstacle tangent line and a road edge in the raster image; the second area is formed by encircling an obstacle tangent and an obstacle outline in the grid map;
the obstacle tangent line is a tangent line of an obstacle curve passing through the position of the current vehicle; the obstacle curve is one or more curves formed by combining obstacle projection points; the obstacle projection point is a part of the projection point of the detection point cloud data on the ground surface plane, which is positioned in the road area;
the obstacle contour is a part close to the position of the current vehicle in two parts formed by dividing the obstacle curve by two obstacle tangent points; the obstacle tangent point is a tangent point of the obstacle tangent line and the obstacle curve.
6. The method for detecting the road edge and the travelable area of the laser radar-based closed road according to claim 1, wherein the step of binarizing and marking the grid map according to the height distribution of the point cloud data detected in the voxel grid comprises the following steps:
calculating the height variance and the three-dimensional surface curvature of the detection point cloud data in the voxel grid;
if the height variance of the detection point cloud data in the voxel grid is greater than a set variance threshold value and the three-dimensional surface curvature of the detection point cloud data in the voxel grid is greater than a set curvature threshold value, marking a grid formed by projection of the voxel grid on a ground surface plane as 1;
and if the height variance of the detection point cloud data in the voxel grid is not greater than a set variance threshold value, or the three-dimensional surface curvature of the detection point cloud data in the voxel grid is not greater than a set curvature threshold value, marking a grid formed by projection of the voxel grid on a ground surface plane as 0.
7. The lidar-based closed road curb and travelable region detection method of claim 6, wherein the three-dimensional surface curvature σ satisfies:
in the formula, λ0、λ1、λ2Is the three eigenvectors of the covariance matrix C, and0<λ1<λ2(ii) a The covariance matrix C satisfies:
in the formula, k is the number of the detection point cloud data in the voxel grid; pi is the coordinate of the ith detection point cloud data in the voxel grid;
and the mean value of the coordinate of the detected point cloud data in the voxel grid is obtained.
8. The lidar based closed road edge and travelable region detection method according to any one of claims 1 to 7, wherein the step of acquiring lidar detection point cloud data comprises:
acquiring original point cloud data of a laser radar;
selecting single-frame point cloud data from the original point cloud data to eliminate ground point cloud and form single-frame detection point cloud data;
and superposing the single-frame detection point cloud data of a set number of adjacent frames to form detection point cloud data.
9. The lidar based closed road edge and drivable area detection method according to claim 8, wherein the step of selecting a single frame of point cloud data from the original point cloud data to exclude ground point clouds and form a single frame of detection point cloud data comprises:
determining a ground surface plane; the ground surface plane is a verification band with the maximum judgment weight; the verification belt is formed by translating a verification plane along the normal direction and is provided with a verification area with a set thickness; the verification plane is obtained by randomly sampling a set number of point cloud fits in single-frame point cloud data; the judgment weight is the proportion of the number of point clouds in the verification band to the number of point clouds outside the verification band;
and eliminating the point cloud data positioned in the ground surface plane from the single-frame point cloud data to form single-frame detection point cloud data.
10. The lidar-based closed road curb and travelable area detection method according to any of claims 1 to 7, wherein the step of determining the curb comprises:
correcting the first path curve and the second path curve according to the curvature of the first path curve and the curvature mean value of the second path curve;
and determining the road edge through the corrected first road edge curve and the second road edge curve.
11. A lidar-based closed road curb and travelable area detection system, comprising:
the data acquisition module is used for acquiring laser radar detection point cloud data;
a voxel grouping module for grouping the detection point cloud data based on voxel grids; the voxel grid is obtained by dividing a set detection space;
the binarization module is used for combining grids formed by projection of the voxel grids on the ground surface plane as a grid map, and binarizing and marking the grid map according to the height distribution of the detection point cloud data in the voxel grids;
and the area determining module is used for determining a road edge and a travelable area according to the grid map of the binarization mark.
12. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor when executing the program implements the steps of the lidar based closed road edge and travelable area detection method according to any of claims 1 to 10.
13. A non-transitory computer readable storage medium having stored thereon a computer program, wherein the computer program when executed by a processor implements the steps of the lidar based closed road edge and travelable area detection method according to any of claims 1 to 10.
14. A computer program product comprising a computer program, wherein the computer program when executed by a processor implements the steps of the lidar based closed road edge and travelable area detection method according to any of claims 1 to 10.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111013260.9A CN113917487A (en) | 2021-08-31 | 2021-08-31 | Laser radar-based method for detecting road edge and drivable area of closed road |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111013260.9A CN113917487A (en) | 2021-08-31 | 2021-08-31 | Laser radar-based method for detecting road edge and drivable area of closed road |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113917487A true CN113917487A (en) | 2022-01-11 |
Family
ID=79233704
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111013260.9A Pending CN113917487A (en) | 2021-08-31 | 2021-08-31 | Laser radar-based method for detecting road edge and drivable area of closed road |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113917487A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114814796A (en) * | 2022-07-01 | 2022-07-29 | 陕西欧卡电子智能科技有限公司 | Method, device and equipment for extracting water surface travelable area based on high-precision map |
| CN115877405A (en) * | 2023-01-31 | 2023-03-31 | 小米汽车科技有限公司 | Method and device for detecting travelable area and vehicle |
| CN116052122A (en) * | 2023-01-28 | 2023-05-02 | 广汽埃安新能源汽车股份有限公司 | Method and device for detecting drivable space, electronic equipment and storage medium |
-
2021
- 2021-08-31 CN CN202111013260.9A patent/CN113917487A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114814796A (en) * | 2022-07-01 | 2022-07-29 | 陕西欧卡电子智能科技有限公司 | Method, device and equipment for extracting water surface travelable area based on high-precision map |
| CN114814796B (en) * | 2022-07-01 | 2022-09-30 | 陕西欧卡电子智能科技有限公司 | Method, device and equipment for extracting water surface travelable area based on high-precision map |
| CN116052122A (en) * | 2023-01-28 | 2023-05-02 | 广汽埃安新能源汽车股份有限公司 | Method and device for detecting drivable space, electronic equipment and storage medium |
| CN115877405A (en) * | 2023-01-31 | 2023-03-31 | 小米汽车科技有限公司 | Method and device for detecting travelable area and vehicle |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111553859B (en) | Laser radar point cloud reflection intensity completion method and system | |
| CN112101066B (en) | Target detection method and device, intelligent driving method and device and storage medium | |
| US11488308B2 (en) | Three-dimensional object detection method and system based on weighted channel features of a point cloud | |
| CN107045629B (en) | Multi-lane line detection method | |
| CN113917487A (en) | Laser radar-based method for detecting road edge and drivable area of closed road | |
| CN109657632B (en) | Lane line detection and identification method | |
| WO2017041396A1 (en) | Driving lane data processing method, device, storage medium and apparatus | |
| CN202134079U (en) | Unmanned vehicle lane marker line identification and alarm device | |
| Zhang et al. | End to end video segmentation for driving: Lane detection for autonomous car | |
| CN112683228A (en) | Monocular camera ranging method and device | |
| Qin et al. | Disparity refinement in depth discontinuity using robustly matched straight lines for digital surface model generation | |
| CN107220632B (en) | Road surface image segmentation method based on normal characteristic | |
| CN114463736A (en) | Multi-target detection method and device based on multi-mode information fusion | |
| Mingzhou et al. | Detection of highway lane lines and drivable regions based on dynamic image enhancement algorithm under unfavorable vision | |
| CN111507340A (en) | Target point cloud data extraction method based on three-dimensional point cloud data | |
| CN113221739B (en) | Monocular vision-based vehicle distance measuring method | |
| CN112733678A (en) | Ranging method, ranging device, computer equipment and storage medium | |
| CN115937825B (en) | Method and device for generating robust lane line under BEV of on-line pitch angle estimation | |
| Chun-Zhao et al. | Drivable road boundary detection for intelligent vehicles based on stereovision with plane-induced homography | |
| KR101910256B1 (en) | Lane Detection Method and System for Camera-based Road Curvature Estimation | |
| Schomerus et al. | Camera-based lane border detection in arbitrarily structured environments | |
| Hirata et al. | Real-time dense depth estimation using semantically-guided LIDAR data propagation and motion stereo | |
| CN114035202A (en) | Obstacle detection method and device based on laser radar | |
| CN112020722B (en) | Three-dimensional sensor data-based road shoulder identification | |
| CN113033395A (en) | Drivable region segmentation method based on DeFCN and vanishing point edge detection |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |