CN114170579A - Road edge detection method and device and automobile - Google Patents
Road edge detection method and device and automobile Download PDFInfo
- Publication number
- CN114170579A CN114170579A CN202010847609.8A CN202010847609A CN114170579A CN 114170579 A CN114170579 A CN 114170579A CN 202010847609 A CN202010847609 A CN 202010847609A CN 114170579 A CN114170579 A CN 114170579A
- Authority
- CN
- China
- Prior art keywords
- points
- point
- road
- seed
- sub
- 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
- 238000003708 edge detection Methods 0.000 title claims abstract description 22
- 238000000034 method Methods 0.000 title claims abstract description 22
- 238000012545 processing Methods 0.000 claims description 6
- 238000000354 decomposition reaction Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 241001632422 Radiola linoides Species 0.000 claims 1
- 238000004364 calculation method Methods 0.000 description 6
- 238000001514 detection method Methods 0.000 description 3
- 241001632427 Radiola Species 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 239000004576 sand Substances 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
Abstract
The invention provides a road edge detection method, a road edge detection device and an automobile, wherein the method comprises the steps of obtaining first point cloud data obtained by scanning a laser radar of the automobile, and converting the first point cloud data into second point cloud data based on a coordinate system of the automobile; dividing an interesting area where the second point cloud data is located into sub-areas at intervals of a preset first distance along the X axis; obtaining seed points of each sub-region based on the second point cloud data in each sub-region by adopting a region growing algorithm, and determining the ground reference height of each sub-region according to part of the seed points or all the seed points; extracting road points in the interested area according to the ground reference height of each sub-area and a preset first height difference; and obtaining a first road edge candidate point from the road surface points in the interested area according to the preset height difference range, the preset gradient range and the normal vector direction of the road edge, and determining the road edge of the road. By the method and the device, the problems of the accuracy and the precision of the conventional road edge detection are solved.
Description
Technical Field
The invention relates to the technical field of automobiles, in particular to a road edge detection method and device and an automobile.
Background
In automatic driving, the detection of the travelable area is an extremely important link, and the road edge is the boundary of the travelable area, so the detection of the road edge plays a key role in the safety and intelligence of automatic driving. However, in reality, urban road scenes are complex, for example, road slopes, both sides of roads, and dynamic and static obstacles in road areas all affect the detection effect of road edges, so that the accuracy and precision of road edge detection are not enough, which is a difficult problem in the prior art.
Disclosure of Invention
The invention aims to provide a road edge detection method, a road edge detection device and an automobile, which are used for solving the problem that the existing road edge detection is insufficient in precision and accuracy.
The invention provides a road edge detection method, which comprises the following steps:
step S11, first point cloud data obtained by scanning of a laser radar of the vehicle are obtained, and the first point cloud data are converted into second point cloud data based on a coordinate system of the vehicle, wherein the X-axis direction of the coordinate system of the vehicle is right in front of the vehicle;
step S12, dividing the region of interest where the second point cloud data is located into sub-regions along the X axis at intervals of a preset first distance;
step S13, obtaining seed points of each sub-region based on the second point cloud data in each sub-region by adopting a region growing algorithm, and determining the ground reference height of each sub-region according to partial seed points or all seed points;
step S14, according to the ground reference height of each sub-area and a preset first height difference, extracting road points in the interested area from the second point cloud data, wherein the road points in the interested area comprise ground points of a road and a road edge;
and step S15, obtaining a first road edge candidate point from the road points in the interested area according to the preset height difference range, the preset gradient range and the normal vector direction of the road edge, and determining the road edge of the road.
Further, the step S13 of obtaining the seed point of each sub-region based on the second point cloud data in the sub-region by using a region growing algorithm specifically includes:
in each sub-area, selecting a preset number of points with the minimum height value in the second point cloud data to calculate the height average value SH；
In each sub-area, selecting the second point cloud data with the height less than SH+SthAs seed points within each of said sub-regions, said SthPresetting a second height difference;
in each sub-region, selecting a seed point to be selected which forms a smooth plane with the seed point from the second point cloud data in the neighborhood of the seed point, taking the seed point to be selected which forms the smooth plane with the seed point as a new seed point, and continuing to select the seed point to be selected which forms the smooth plane with the new seed point from the second point cloud data in the neighborhood of the new seed point until the boundary of the region of interest is reached, or reaching the condition: and the neighborhood of all the seed points does not have the seed points to be selected which form a smooth plane with all the seed points.
Further, in each sub-region, selecting a seed point to be selected in the second point cloud data in the neighborhood of the seed point, where the seed point and the seed point form a smooth plane, and using the seed point to be selected in the neighborhood of the seed point and the seed point form a smooth plane as a new seed point, specifically including:
in each sub-region, taking the second point cloud data in the neighborhood of the seed point as a seed point to be selected;
judging whether an included angle between the seed point and a normal vector of the seed point to be selected is smaller than a preset angle threshold value or not;
and when the included angle between the seed point and the normal vector of the seed point to be selected is smaller than a preset angle threshold value, judging that the seed point to be selected and the seed point form a smooth plane, and determining that the seed point to be selected is a new seed point.
Further, the step S13 of determining the ground reference height of each sub-area according to a part of the seed points or all the seed points specifically includes:
according to whatCalculating the height average value of the partial seed points or the height average value R of all the seed pointsHSaid R isHA ground reference height for each of said sub-areas.
Further, step S14 specifically includes:
and in each sub-area, extracting a point with the height lower than the sum of the ground reference height and a preset second height difference from the second point cloud data as a road point in the interested area.
Further, step S15 specifically includes:
rasterizing the road points in the interested area, and establishing a three-dimensional grid map according to preset length, width and height;
calculating the maximum height difference of the pavement points of each grid, and judging whether the maximum height difference of the pavement points of each grid conforms to a preset height difference range or not;
calculating the height average value of all road surface points in each grid by using a formula
Calculating the gradient of each grid, and judging whether the gradient of each grid meets a preset gradient range or not; wherein G is the gradient of each grid, and I is the height average value of all road points of each grid;
using formulas
Calculating the mean value of all the road surface points of each grid, wherein the mean value is obtained
Is the mean of all road points in each grid, | Vi,j,kI denotes the grid Vi,j,kThe number of all road points in, said plIs a grid Vi,j,kThe 3D coordinates of the ith road point;
using formulas
Calculating each grid covariance, wherein Ci,j,kFor covariance, the T represents the transpose of the matrix;
singular value decomposition is carried out on the covariance of each grid, a singular vector with the minimum singular value is selected as a normal vector of each grid, and whether the normal vector of each grid is perpendicular to the vehicle is judged;
when the grids simultaneously satisfy: the maximum height difference accords with a preset height difference range, the gradient accords with a preset gradient range, the normal vector is perpendicular to the vehicle, the ground point in the grid is determined as a first road edge candidate point, and the road edge of the road is determined.
Further, the step S15 further includes:
deleting the first path candidate point in the non-continuity grid as an isolated noise point to obtain a second path candidate point;
and determining the second road edge candidate point as the road edge of the road.
Further, the step S15 further includes:
removing obstacle noise in the second path of edge candidate points according to an obstacle clustering algorithm to obtain third path of edge candidate points;
and determining the third route candidate point as the route of the road.
The invention provides a road edge detection device, which comprises:
the system comprises an acquisition unit, a processing unit and a processing unit, wherein the acquisition unit is used for acquiring first point cloud data obtained by scanning a laser radar of a vehicle and converting the first point cloud data into second point cloud data based on a coordinate system of the vehicle, and the X-axis direction of the coordinate system of the vehicle is right ahead of the vehicle;
the dividing unit is used for dividing the region of interest where the second point cloud data is located into sub-regions along the X axis at intervals of a preset first distance;
the computing unit is used for obtaining the seed points of each sub-region based on the second point cloud data in each sub-region by adopting a region growing algorithm, and determining the ground reference height of each sub-region according to part of the seed points or all the seed points;
the extracting unit is used for extracting road points in the interested area from the second point cloud data according to the ground reference height of each sub-area and a preset first height difference, wherein the road points in the interested area comprise road surfaces and road edge ground points;
and the determining unit is used for obtaining a first road edge candidate point from the road points in the interested area according to the preset height difference range, the preset gradient range and the normal vector direction of the road edge, and determining the road edge of the road.
The invention provides an automobile which comprises the road edge detection device.
The implementation of the invention has the following beneficial effects:
according to the method, point cloud data are obtained through scanning of a laser radar, the point cloud data are converted into a vehicle coordinate system, sub-areas are divided into along the front of a vehicle according to a preset first distance, seed points are selected from the sub-areas to grow, the ground reference height of each sub-area is determined according to the seed points, ground points of a road surface and the road edge are screened out and subjected to rasterization processing through the road surface reference height of each sub-area and the preset first height difference, the road edge is preliminarily screened out through the height difference, the gradient and the normal vector direction between the road edge and the road surface, and road surface interference and obstacle interference are filtered, so that the problem that the existing road edge detection accuracy and accuracy are not enough is solved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a flowchart of a method for detecting a road edge according to an embodiment of the present invention.
Fig. 2 is a front view of an automobile provided by an embodiment of the present invention.
Fig. 3 is a left side view of an automobile provided by an embodiment of the present invention.
Fig. 4 is a road surface map for dividing sub-regions according to an embodiment of the present invention.
Fig. 5 is a structural diagram of a road edge detection device according to an embodiment of the present invention.
Detailed Description
In this patent, the road edges are gradually selected by geometric features and interference elimination, and the following description will further explain this embodiment with reference to the drawings and examples.
As shown in fig. 1, an embodiment of the present invention provides a road edge detection method, where the method includes:
and S11, acquiring first point cloud data obtained by scanning the laser radar of the vehicle, and converting the first point cloud data into second point cloud data based on a coordinate system of the vehicle, wherein the X-axis direction of the coordinate system of the vehicle is right ahead of the vehicle.
Referring to FIG. 2, ysAnd zsRespectively Y-axis coordinate and Z-axis coordinate of laser radar coordinate system, YvAnd zvRespectively a Y-axis coordinate and a Z-axis coordinate of a coordinate system of the vehicle; referring to FIG. 3, xsAnd zsRespectively X-axis coordinate and Z-axis coordinate of laser radar coordinate system, XvAnd zvThe X-axis coordinate and the Z-axis coordinate of the vehicle coordinate system are respectively, the coordinate of the laser radar coordinate system can be converted into the coordinate of the vehicle coordinate system, and the X-axis direction of the vehicle coordinate system is the right front of the vehicle.
It should be noted that, the conversion of the first point cloud data into the second point cloud data is for convenience of calculation and testing.
And step S12, dividing the region of interest where the second point cloud data is located into sub-regions along the X axis at intervals of a preset first distance.
It should be noted that, in order to improve the calculation efficiency, the region of interest is selected for the second point cloud data, and x ∈ [ x ] will be satisfiedmin,xmax]and y∈[ymin,ymax]and z∈[zmin,zmax]The region is taken as the region of interest, xminX is the same asmaxRespectively a preset X-axis minimum coordinate and a preset X-axis maximum coordinate, yminThe ymaxRespectively a preset Y-axis minimum coordinate and a preset Y-axis maximum coordinate, the zminZ tomaxRespectively setting a preset Z-axis minimum coordinate and a preset Z-axis maximum coordinate; since the coordinates are obtained by lidar scanning of the vehicle, x isminX is the same asmaxAll in the positive X-axis direction.
Referring to fig. 4, the road edge 41 and the road edge 42 are respectively located at two sides of the road, the boundary 43 and the boundary 44 are two side boundaries of the region of interest where the second point cloud data is located along the X-axis, and the plurality of dividing lines 45 are divided into sub-regions at intervals of a preset first distance along the X-axis of the coordinate system of the vehicle.
And step S13, obtaining seed points of each sub-region based on the second point cloud data in each sub-region by adopting a region growing algorithm, and determining the ground reference height of each sub-region according to partial seed points or all seed points.
Specifically, obtaining the seed point of each sub-region based on the second point cloud data in the sub-region by using a region growing algorithm includes:
in each sub-area, selecting a preset number of points with the minimum height value in the second point cloud data to calculate the height average value SH；
In each sub-area, selecting the second point cloud data with the height less than SH+SthAs seed points within each of said sub-regions, said SthPresetting a second height difference;
it should be noted that, preferably, in each sub-region, if the number of seed points is less than the preset number, the average height value S of the lower points in each sub-region is calculatedHThen, averaging the preset number of points with the second minimum height to calculate, and excluding the point with the minimum height; selecting the height smaller than S in each sub-areaH+SthBy makingFor seed points in each of said sub-regions, said SthIn order to preset the second height difference, the average height of the lower points in the sub-regions is increased to increase the number of seed points, so as to avoid interference caused by the fact that the heights of individual points in some sub-regions are extremely low.
In each sub-region, selecting a seed point to be selected which forms a smooth plane with the seed point from the second point cloud data in the neighborhood of the seed point, taking the seed point to be selected which forms the smooth plane with the seed point as a new seed point, and continuing to select the seed point to be selected which forms the smooth plane with the new seed point from the second point cloud data in the neighborhood of the new seed point until the boundary of the region of interest is reached, or reaching the condition: and the neighborhood of all the seed points does not have the seed points to be selected which form a smooth plane with all the seed points.
It should be noted that the neighborhood is relative to the seed point, and means that the distance to the seed point does not exceed a set of preset positive numbers, a new seed point is found in the neighborhood of the seed point, and then the new seed point is continuously used to find the seed point in the neighborhood of the new seed point until the boundary of the region of interest is reached or no point forming a smooth plane with all the seed points is found in the neighborhood of all the seed points, so as to find out all the seed points in each sub-region.
Further, in each sub-region, selecting a seed point to be selected in the second point cloud data in the neighborhood of the seed point, where the seed point and the seed point form a smooth plane, and using the seed point to be selected in the neighborhood of the seed point and the seed point form a smooth plane as a new seed point, specifically including:
in each sub-region, taking the second point cloud data in the neighborhood of the seed point as a seed point to be selected;
judging whether an included angle between the seed point and a normal vector of the seed point to be selected is smaller than a preset angle threshold value or not;
and when the included angle between the seed point and the normal vector of the seed point to be selected is smaller than a preset angle threshold value, judging that the seed point to be selected and the seed point form a smooth plane, and determining that the seed point to be selected is a new seed point.
Further, the step S13 of determining the ground reference height of each sub-area according to a part of the seed points or all the seed points specifically includes:
calculating the height average value of the partial seed points or the height average value R of all the seed points according to the partial seed points or all the seed pointsHSaid R isHA ground reference height for each of said sub-areas.
It should be noted that, the average calculation is performed according to part of the seed points or all the seed points in each sub-region, and the ground reference height of each sub-region is determined; the average calculation using the partial seed points is based on the number of the preset seed points and the average calculation using the partial seed points with the lowest height.
Step S14, according to the ground reference height of each sub-area and a preset first height difference, extracting the road points in the interested area from the second point cloud data, wherein the road points in the interested area comprise the ground points of the road surface and the road edge.
Specifically, step S14 specifically includes:
and in each sub-area, extracting a point with the height lower than the sum of the ground reference height and a preset first height difference from the second point cloud data as a road point in the interested area.
It should be noted that, in step S14, the ground reference height R of each sub-area is obtained through calculationHThe preset second height difference is RthThe sum of the ground reference height and the preset first height difference is RH+RthEach sub-area is lower than the point of the sum of the ground reference height and a preset first height difference, and the point is extracted as a road point in the interested area, wherein the preset first height difference is RthThe general curb height is taken into account.
And step S15, obtaining a first road edge candidate point from the road points in the interested area according to the preset height difference range, the preset gradient range and the normal vector direction of the road edge, and determining the road edge of the road.
Step S15 specifically includes:
rasterizing the road points in the interested area, and establishing a three-dimensional grid map according to preset length, width and height;
it should be noted that, the above steps are actually to establish a 3D grid map for the point cloud data in the coordinate system of the vehicle body itself, and the size of each grid is set to be Vl*Vw*VhIn which V isl、Vw、VhRespectively representing the length, width and height of the grid.
Calculating the maximum height difference of the pavement points of each grid, and judging whether the maximum height difference of the pavement points of each grid conforms to a preset height difference range or not;
it should be noted that the maximum height difference of each grid of road surface points refers to the maximum height difference of two road surface points in each grid, and the road surface points include both ground points of the road surface and ground points of the road edge; generally, at the road edge, the height difference range between the ground point of the road edge and the ground point of the road surface is higher than that between the ground points of the common road surface, and the height difference range between the ground point of the road edge and the ground point of the road surface is less than 15cm or 20 cm.
Calculating the height average value of all road surface points in each grid by using a formula
Calculating the gradient of each grid, and judging whether the gradient of each grid meets a preset gradient range or not; wherein G is the gradient of each grid, and I is the height average value of all road points of each grid;
the road surface points include road edge ground points and road surface ground points.
Using formulas
Calculating the mean value of all the road surface points of each grid, wherein the mean value is obtained
Is the mean of all road points of each grid, | Vi,j,kI denotes the grid Vi,j,kThe number of all road points in, said plIs a grid Vi,j,kThe 3D coordinates of the ith road point;
using formulas
Calculating each grid covariance, wherein Ci,j,kFor covariance, the T represents the transpose of the matrix;
singular value decomposition is carried out on the covariance of each grid, a singular vector with the minimum singular value is selected as a normal vector of each grid, and whether the normal vector of each grid is perpendicular to the vehicle is judged;
when the grids simultaneously satisfy: the maximum height difference accords with a preset height difference range, the gradient accords with a preset gradient range, the normal vector is perpendicular to the vehicle, the road surface point in the grid is determined as a first road edge candidate point, and the first road edge candidate point is determined as the road edge of the road.
It should be noted that the first candidate point of the road edge also has noise and interference, such as discontinuous portions, some road surfaces have convex portions, and some road surface areas have obstacles, which all need to be eliminated step by step to obtain a more accurate road edge.
Further, the step S15 further includes:
deleting the first path candidate point in the grid without continuity as an isolated noise point to obtain a second path candidate point,
and determining the second road edge candidate point as the road edge of the road.
It should be noted that the screening step is based on continuity of the grid included in the road edge.
Specifically, clustering is carried out through the distance between each grid and the adjacent grids around, if the number of grids in the same cluster is less than the preset number of grids, the grids are determined not to have continuity, and the grids belong to interference information to be excluded.
Further, the step S15 further includes:
removing obstacle noise in the second path of edge candidate points according to an obstacle clustering algorithm to obtain third path of edge candidate points;
and determining the third route candidate point as the route of the road.
Specifically, since the selection of the road edge is easily affected by noise, such as other vehicles in the road, road signs, and the like, the points located in the area of the obstacle are removed from the road edge candidate points by clustering the divided non-ground points according to the distance and then projecting them into the ground. Therefore, noise generated by the similarity of the local point cloud characteristics of the obstacles and the road edge characteristics can be reduced; the third path candidate point is the result required by the present invention.
As shown in fig. 5, an embodiment of the present invention provides a road edge detection apparatus, including:
the acquisition unit 51 is used for acquiring first point cloud data obtained by scanning a laser radar of a vehicle and converting the first point cloud data into second point cloud data based on a coordinate system of the vehicle, wherein the X-axis direction of the coordinate system of the vehicle is right ahead of the vehicle;
the dividing unit 52 is configured to divide the region of interest where the second point cloud data is located into sub-regions along the X axis at intervals of a preset first distance;
the calculating unit 53 is configured to obtain seed points of each sub-region based on the second point cloud data in each sub-region by using a region growing algorithm, and determine a ground reference height of each sub-region according to part of the seed points or all the seed points;
an extracting unit 54, configured to extract, from the second point cloud data, road points in the interested area according to a ground reference height of each sub-area and a preset first height difference, where the road points in the interested area include ground points of a road surface and a road edge;
and the determining unit 55 is configured to obtain a first road edge candidate point from the road points in the interested area according to the preset height difference range, the preset gradient range and the normal vector direction of the road edge, and determine the road edge of the road.
The invention also provides an automobile which comprises the road edge detection device.
The implementation of the invention has the following beneficial effects:
according to the method, point cloud data are obtained through scanning of a laser radar, the point cloud data are converted into a vehicle coordinate system, the point cloud data are divided into sub-areas along an X axis of the vehicle coordinate system according to a preset first distance, seed points are selected from the sub-areas to grow, the ground reference height of each sub-area is determined according to the seed points, ground points of a road surface and the road edge are screened out and subjected to rasterization processing through the road surface reference height of each sub-area and a preset second height difference, the road edge is preliminarily screened out through the height difference, gradient and normal vector directions between the road edge and the road surface, and road surface interference and obstacle interference are filtered, so that the problem that the existing road edge detection accuracy and accuracy are not enough is solved.
The foregoing is a more detailed description of the invention in connection with specific preferred embodiments and it is not intended that the invention be limited to these specific details. For those skilled in the art to which the invention pertains, several simple deductions or substitutions can be made without departing from the spirit of the invention, and all shall be considered as belonging to the protection scope of the invention.
Claims (10)
1. A method of road edge detection, the method comprising:
step S11, first point cloud data obtained by scanning of a laser radar of the vehicle are obtained, and the first point cloud data are converted into second point cloud data based on a coordinate system of the vehicle, wherein the X-axis direction of the coordinate system of the vehicle is right in front of the vehicle;
step S12, dividing the region of interest where the second point cloud data is located into sub-regions along the X axis at intervals of a preset first distance;
step S13, obtaining seed points of each sub-region based on the second point cloud data in each sub-region by adopting a region growing algorithm, and determining the ground reference height of each sub-region according to partial seed points or all seed points;
step S14, extracting road points in the interested area from the second point cloud data according to the ground reference height of each sub-area and a preset first height difference, wherein the road points in the interested area comprise the ground points of a road surface and a road edge;
and step S15, obtaining a first road edge candidate point from the road points in the interested area according to the preset height difference range, the preset gradient range and the normal vector direction of the road edge, and determining the road edge of the road.
2. The method of claim 1, wherein the step S13 of obtaining the seed point of each sub-region based on the second point cloud data of the sub-region by using a region growing algorithm specifically comprises:
in each sub-area, selecting a preset number of points with the minimum height value in the second point cloud data to calculate the height average value SH；
In each sub-area, selecting the second point cloud data with the height less than SH+SthAs seed points within each of said sub-regions, said SthPresetting a second height difference;
in each sub-region, selecting a seed point to be selected which forms a smooth plane with the seed point from the second point cloud data in the neighborhood of the seed point, taking the seed point to be selected which forms the smooth plane with the seed point as a new seed point, and continuing to select the seed point to be selected which forms the smooth plane with the new seed point from the second point cloud data in the neighborhood of the new seed point until the boundary of the region of interest is reached, or reaching the condition: and the neighborhood of all the seed points does not have the seed points to be selected which form a smooth plane with all the seed points.
3. The method of claim 2, wherein, in each sub-region, selecting candidate seed points that form a smooth plane with the seed point from the second point cloud data in the neighborhood of the seed point, and taking the candidate seed points that form the smooth plane with the seed point as new seed points, specifically comprises:
in each sub-region, taking the second point cloud data in the neighborhood of the seed point as a seed point to be selected;
judging whether an included angle between the seed point and a normal vector of the seed point to be selected is smaller than a preset angle threshold value or not;
and when the included angle between the seed point and the normal vector of the seed point to be selected is smaller than a preset angle threshold value, judging that the seed point to be selected and the seed point form a smooth plane, and determining that the seed point to be selected is a new seed point.
4. The method according to claim 1, wherein the step S13 of determining the ground reference height of each sub-area according to a part of the seed points or all the seed points specifically comprises:
calculating the height average value of the partial seed points or the height average value R of all the seed points according to the partial seed points or all the seed pointsHSaid R isHA ground reference height for each of said sub-areas.
5. The method according to claim 1, wherein step S14 specifically includes:
and in each sub-area, extracting a point with the height lower than the sum of the ground reference height and a preset first height difference from the second point cloud data as a road point in the interested area.
6. The method according to claim 1, wherein step S15 specifically includes:
rasterizing the road points in the interested area, and establishing a three-dimensional grid map according to preset length, width and height;
calculating the maximum height difference of the pavement points of each grid, and judging whether the maximum height difference of the pavement points of each grid conforms to a preset height difference range or not;
calculating the height average value of all road surface points in each grid by using a formula
Calculating the gradient of each grid, and judging whether the gradient of each grid meets a preset gradient range or not; wherein G is the gradient of each grid, and I is the height average value of all road points of each grid;
using formulas
Calculating the mean value of all the road surface points of each grid, wherein the mean value is obtained
Is the mean of all road points in each grid, | Vi,j,kI denotes the grid Vi,j,kThe number of all road points in, said plIs a grid Vi,j,k3D coordinates of the ith road point;
using formulas
Calculating each grid covariance, wherein Ci,j,kFor covariance, the T represents the transpose of the matrix;
singular value decomposition is carried out on the covariance of each grid, a singular vector with the minimum singular value is selected as a normal vector of each grid, and whether the normal vector of each grid is perpendicular to the vehicle is judged;
when the grids simultaneously satisfy: and if the maximum height difference conforms to a preset height difference range, the gradient conforms to a preset gradient range and the normal vector is perpendicular to the vehicle, determining the road surface point in the grid as a first road edge candidate point and determining the first road edge candidate point as the road edge of the road.
7. The method according to claim 6, wherein the step S15 further comprises:
deleting the first path candidate point in the grid without continuity as an isolated noise point to obtain a second path candidate point;
and determining the second road edge candidate point as the road edge of the road.
8. The method of claim 7, wherein the step S15 further comprises:
removing obstacle noise in the second path of edge candidate points according to an obstacle clustering algorithm to obtain third path of edge candidate points;
and determining the third route candidate point as the route of the road.
9. A road edge detection device, the device comprising:
the system comprises an acquisition unit, a processing unit and a processing unit, wherein the acquisition unit is used for acquiring first point cloud data obtained by scanning a laser radar of a vehicle and converting the first point cloud data into second point cloud data based on a coordinate system of the vehicle, and the X-axis direction of the coordinate system of the vehicle is right ahead of the vehicle;
the dividing unit is used for dividing the region of interest where the second point cloud data is located into sub-regions along the X axis at intervals of a preset first distance;
the computing unit is used for obtaining the seed points of each sub-region based on the second point cloud data in each sub-region by adopting a region growing algorithm, and determining the ground reference height of each sub-region according to part of the seed points or all the seed points;
the extracting unit is used for extracting road points in the interested area from the second point cloud data according to the ground reference height of each sub-area and a preset first height difference, wherein the road points in the interested area comprise the ground points of a road surface and a road edge;
and the determining unit is used for obtaining a first road edge candidate point from the road points in the interested area according to the preset height difference range, the preset gradient range and the normal vector direction of the road edge, and determining the road edge of the road.
10. An automobile, characterized in that the automobile comprises the road edge detection device according to claim 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010847609.8A CN114170579A (en) | 2020-08-21 | 2020-08-21 | Road edge detection method and device and automobile |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010847609.8A CN114170579A (en) | 2020-08-21 | 2020-08-21 | Road edge detection method and device and automobile |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114170579A true CN114170579A (en) | 2022-03-11 |
Family
ID=80475386
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010847609.8A Pending CN114170579A (en) | 2020-08-21 | 2020-08-21 | Road edge detection method and device and automobile |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114170579A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2023226147A1 (en) * | 2022-05-26 | 2023-11-30 | 惠州市德赛西威汽车电子股份有限公司 | Curb detection method and apparatus for automatic parking, vehicle, and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107272019A (en) * | 2017-05-09 | 2017-10-20 | 深圳市速腾聚创科技有限公司 | Curb detection method based on Laser Radar Scanning |
| WO2018205119A1 (en) * | 2017-05-09 | 2018-11-15 | 深圳市速腾聚创科技有限公司 | Roadside detection method and system based on laser radar scanning |
| CN109738910A (en) * | 2019-01-28 | 2019-05-10 | 重庆邮电大学 | A kind of curb detection method based on three-dimensional laser radar |
| CN109858460A (en) * | 2019-02-20 | 2019-06-07 | 重庆邮电大学 | A kind of method for detecting lane lines based on three-dimensional laser radar |
| CN110781827A (en) * | 2019-10-25 | 2020-02-11 | 中山大学 | Road edge detection system and method based on laser radar and fan-shaped space division |
-
2020
- 2020-08-21 CN CN202010847609.8A patent/CN114170579A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107272019A (en) * | 2017-05-09 | 2017-10-20 | 深圳市速腾聚创科技有限公司 | Curb detection method based on Laser Radar Scanning |
| WO2018205119A1 (en) * | 2017-05-09 | 2018-11-15 | 深圳市速腾聚创科技有限公司 | Roadside detection method and system based on laser radar scanning |
| CN109738910A (en) * | 2019-01-28 | 2019-05-10 | 重庆邮电大学 | A kind of curb detection method based on three-dimensional laser radar |
| CN109858460A (en) * | 2019-02-20 | 2019-06-07 | 重庆邮电大学 | A kind of method for detecting lane lines based on three-dimensional laser radar |
| CN110781827A (en) * | 2019-10-25 | 2020-02-11 | 中山大学 | Road edge detection system and method based on laser radar and fan-shaped space division |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2023226147A1 (en) * | 2022-05-26 | 2023-11-30 | 惠州市德赛西威汽车电子股份有限公司 | Curb detection method and apparatus for automatic parking, vehicle, and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109858460B (en) | Lane line detection method based on three-dimensional laser radar | |
| CN110443225B (en) | Virtual and real lane line identification method and device based on feature pixel statistics | |
| CN111079611B (en) | Automatic extraction method for road surface and marking line thereof | |
| CN111325138B (en) | Road boundary real-time detection method based on point cloud local concave-convex characteristics | |
| CN111179152B (en) | Road identification recognition method and device, medium and terminal | |
| CN103714538B (en) | road edge detection method, device and vehicle | |
| CN111537994B (en) | Unmanned mine card obstacle detection method | |
| CN108845569A (en) | Generate semi-automatic cloud method of the horizontal bend lane of three-dimensional high-definition mileage chart | |
| CN108898672A (en) | A kind of semi-automatic cloud method making three-dimensional high-definition mileage chart lane line | |
| CN112200171B (en) | Road point cloud extraction method based on scanning lines | |
| CN112801022A (en) | Method for rapidly detecting and updating road boundary of unmanned mine card operation area | |
| CN110197173B (en) | Road edge detection method based on binocular vision | |
| CN112561944A (en) | Lane line extraction method based on vehicle-mounted laser point cloud | |
| CN112001216A (en) | Automobile driving lane detection system based on computer | |
| CN110532963B (en) | Vehicle-mounted laser radar point cloud driven road marking accurate extraction method | |
| CN105678287A (en) | Ridge-measure-based lane line detection method | |
| CN112014856A (en) | Road edge extraction method and device suitable for cross road section | |
| CN115451983A (en) | Dynamic environment mapping and path planning method and device in complex scene | |
| CN114877838B (en) | Road geometric feature detection method based on vehicle-mounted laser scanning system | |
| CN112070756A (en) | Three-dimensional road surface disease measuring method based on unmanned aerial vehicle oblique photography | |
| CN107220632B (en) | Road surface image segmentation method based on normal characteristic | |
| CN114325755A (en) | Retaining wall detection method and system suitable for automatic driving vehicle | |
| CN117073664B (en) | Surface mine road topography mapping method | |
| CN114170579A (en) | Road edge detection method and device and automobile | |
| CN116188334B (en) | Automatic repair method and device for lane line point cloud |
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 |