CN115761682A - Method and device for identifying travelable area based on laser perception and intelligent mine card - Google Patents
Method and device for identifying travelable area based on laser perception and intelligent mine card Download PDFInfo
- Publication number
- CN115761682A CN115761682A CN202211531348.4A CN202211531348A CN115761682A CN 115761682 A CN115761682 A CN 115761682A CN 202211531348 A CN202211531348 A CN 202211531348A CN 115761682 A CN115761682 A CN 115761682A
- Authority
- CN
- China
- Prior art keywords
- points
- point cloud
- point
- boundary
- ground
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Traffic Control Systems (AREA)
Abstract
The invention belongs to the technical field of automatic driving perception, and particularly provides a method and a device for identifying a travelable area based on laser perception, and an intelligent mine card, wherein the method comprises the following steps: converting the point cloud from a radar coordinate system to a vehicle body coordinate system and preprocessing point cloud data; performing ground plane fitting by using the preprocessed point cloud, and segmenting ground points from the preprocessed point cloud; dividing the divided ground points by using a rectangular network with a set interval, finding out a boundary grid and extracting boundary points; performing boundary line fitting on the extracted boundary points, filling the boundary points of the vacant positions, and eliminating points with abnormal value deviation larger than a set threshold value; and issuing road boundary points identified based on the laser point cloud, and issuing travelable areas of the mine car. An effective method is provided for identifying the travelable area of the unstructured road in the mountainous area in the current automatic driving field.
Description
Technical Field
The invention belongs to the technical field of automatic driving perception, and particularly relates to a drivable area detection method and device for a mine unstructured road, which are used for real-time output of a mine area road virtual boundary line and a drivable area through processing of laser radar point cloud and are used for decision, planning and control of an automatic driving mine car, and particularly relates to a drivable area identification method and device based on laser perception and an intelligent mine card.
Background
The mine road is rugged and winding, the working environment is severe, the traffic accident occurrence rate is high, and along with the rapid development of the automatic driving technology, the intellectualization and the humanization of the mining area gradually become the development direction of the mine. The construction and development of the intelligent mine can reduce errors caused by manual operation and improve the safety of mining operation, and has great significance for mining industry exploitation in China. The automatic driving mine car can improve the utilization rate of equipment, reduce the labor cost and improve the production efficiency of a mining area; meanwhile, by continuously optimizing the control algorithm, the oil consumption can be reduced, the loss of the mine car equipment is reduced, and the service life of the mine car equipment is prolonged.
The sensing technology of the automatic driving mine card can enable the mine car to better understand the surrounding environment, the position, the distance and the speed of the obstacle, the driving area and the like, so that the decision, the planning and the control functions of the rear end are better realized. The invention mainly aims at sensing the boundary of a mine road and a travelable area, defines a safe driving range for an automatic driving mine car and ensures the safety of the mine car in the transportation process. The sensor included in the automatic driving system sensing technology comprises a millimeter wave radar, a camera, a laser radar and the like, and the sensing mode based on the laser radar is mainly adopted. On one hand, the laser radar can be used in all weather without being influenced by illumination and weather change conditions such as rainy days; on the other hand, the laser radar can obtain point cloud data with high precision, high latitude and wide visual field range, and is more suitable for detecting travelable areas of mine roads.
At present, mine road detection has a plurality of technical difficulties, mine roads have no obvious road boundaries, some places are heaves piled by stones, some places have the same flatness as the ground, and single-lane roads and double-lane roads are mixed. In addition, the mining area road always has a certain gradient, and some potholes and bulges also exist in the middle of the road, which brings certain challenges to the detection of the algorithm.
Disclosure of Invention
Mine road detection has a plurality of technical difficulties, mine roads have no obvious road boundaries, some places are bumps piled by stones, some places have the same flatness as the ground, and one-way roads and two-way roads are mixed. And moreover, the mine road always has a certain gradient, and the middle of the road also has some hollows and bulges, which bring certain challenges to the detection of the algorithm. The invention discloses a method and a device for identifying a travelable area based on laser sensing and an intelligent mine card.
In a first aspect, the technical solution of the present invention provides a method for identifying a travelable area based on laser sensing, including the steps of:
converting the point cloud from a radar coordinate system to a vehicle body coordinate system and preprocessing point cloud data;
performing ground plane fitting by using the preprocessed point cloud, and segmenting ground points from the preprocessed point cloud;
dividing the divided ground points by using a rectangular network with set intervals, finding out a boundary grid and extracting boundary points;
performing boundary line fitting on the extracted boundary points, filling the boundary points of the vacant positions, and eliminating points with abnormal value deviation larger than a set threshold value;
and issuing road boundary points identified based on the laser point cloud, and issuing travelable areas of the mine car.
As a further limitation of the technical solution of the present invention, the step of converting the point cloud from the radar coordinate system to the vehicle body coordinate system and preprocessing the point cloud data includes:
acquiring front road point cloud data acquired by a laser radar, and performing point cloud downsampling filtering;
traversing all points in the point cloud, eliminating points with coordinate values equal to the NaN value and eliminating points with coordinate values smaller than a first threshold value;
carrying out coordinate system transformation on the point cloud data, and transferring the point cloud from a radar coordinate system to a vehicle body coordinate system;
and (4) defining a plane point cloud to detect the ROI, and intercepting the point cloud positioned in the ROI to obtain a preprocessed point cloud.
As a further limitation of the technical scheme of the invention, the step of using the preprocessed point cloud to perform ground plane fitting and dividing ground points from the preprocessed point cloud comprises the following steps:
sorting the preprocessed point clouds according to the height of the z axis;
calculating the lowest point of the point cloud, and selecting the point of the point cloud with the height difference between the Z-axis height and the lowest point within a second threshold value as a seed point to generate a seed point set;
establishing an initial plane model for describing the ground according to the seed points;
calculating the distance of the orthogonal projection of each point in the point cloud to the plane, and comparing the distance with a set threshold Th dist Comparing when the height difference is less than threshold Th dist Considering the point as belonging to the ground point, when the height difference is greater than the threshold Th dist If the point is not the ground point;
all classified ground points are taken as a seed point set of next iteration, and iteration optimization is continued; and (5) obtaining the iteration times, and finally obtaining the segmented ground point cloud.
As a further limitation of the technical solution of the present invention, the step of establishing an initial plane model describing the ground according to the seed points comprises:
and evaluating the established initial plane model.
As a further limitation of the technical solution of the present invention, the step of segmenting the segmented ground points by using a rectangular network with a set interval, finding a boundary mesh and extracting boundary points comprises:
receiving the segmented ground point cloud;
converting the 3D data point set to an xy plane through coordinate conversion to obtain a plane point set;
establishing a minimum bounding box of the data point set; and the ground point cloud is divided by using a rectangular network with set intervals, a boundary grid is found out, and boundary points are extracted.
As a further limitation of the technical scheme of the invention, a minimum bounding box of the data point set is established; the method comprises the following steps of dividing the ground point cloud by using a rectangular network with set intervals, finding out a boundary grid and extracting boundary points:
traversing all points of the plane point set to obtain the maximum value and the minimum value of the X axis and the maximum value and the minimum value of the Y axis and establishing a minimum bounding box of the point set;
and putting the data points into corresponding network units, finding out a boundary grid and extracting boundary points.
As a further limitation of the technical solution of the present invention, the step of placing the data points into the corresponding network units, finding the boundary grid and extracting the boundary points comprises:
after the ground point cloud is gridded, calculating the row and the column where the point is located according to the coordinate of the current point, adding the point into the grid corresponding to the row and the column, wherein if the number of points in the corresponding grid is zero, the pixel value is 0, and otherwise, the pixel value is 1;
judging each grid, if the number of points of the grid is not zero and the number of points in the adjacent grid is not zero, determining the grid as a non-boundary grid, otherwise determining the grid as a boundary grid;
and putting the point cloud in the boundary grid into a corresponding storage container to obtain the road boundary points.
As a further limitation of the technical scheme of the invention, the steps of performing boundary line fitting on the extracted boundary points, filling the boundary points of the vacant positions, and eliminating the points with abnormal value deviation larger than a set threshold value comprise:
and performing segmentation fitting on the obtained boundary points by using a B spline curve, filling the boundary points at the vacant positions, and removing part of abnormal points.
In a second aspect, the technical solution of the present invention further provides a device for identifying a travelable region based on laser sensing, which includes a preprocessing module, a ground point segmentation module, a boundary point acquisition module, a boundary line fitting module, and a publishing module;
the preprocessing module is used for converting the point cloud from a radar coordinate system to a vehicle body coordinate system and preprocessing the point cloud data;
the ground point segmentation module is used for performing ground plane fitting by using the preprocessed point cloud and segmenting ground points from the preprocessed point cloud;
the boundary point acquisition module is used for segmenting the segmented ground points by using a rectangular network with set intervals, finding out a boundary grid and extracting boundary points;
the boundary line fitting module is used for performing boundary line fitting on the extracted boundary points, filling the boundary points of the vacant positions and eliminating points with abnormal value deviation larger than a set threshold value;
and the issuing module is used for issuing road boundary points identified based on the laser point cloud and issuing travelable areas of the mine car.
As a further limitation of the technical solution of the present invention, the preprocessing module includes a sampling filtering unit, a first preprocessing unit, a coordinate conversion unit, and a second preprocessing unit;
the sampling filtering unit is used for acquiring front road point cloud data acquired by the laser radar and performing point cloud downsampling filtering;
the first preprocessing unit is used for traversing all points in the point cloud, eliminating points with coordinate values equal to a NaN value and eliminating points with coordinate values smaller than a first threshold value;
the coordinate conversion unit is used for carrying out coordinate system conversion on the point cloud data and converting the point cloud from the radar coordinate system to the vehicle body coordinate system;
and the second preprocessing unit is used for defining a plane point cloud detection ROI, intercepting the point cloud positioned in the ROI and obtaining a preprocessed point cloud.
As a further limitation of the technical scheme of the invention, the ground point segmentation module comprises a sorting unit, a generating unit, a plane establishing unit, a calculation processing unit and an iteration segmentation unit;
the sorting unit is used for sorting the preprocessed point cloud according to the height of the z axis;
the generating unit is used for calculating the lowest point of the point cloud, and selecting the point of the point cloud with the height difference between the Z-axis height and the lowest point within a second threshold value as a seed point to generate a seed point set;
the plane establishing unit is used for establishing an initial plane model describing the ground according to the seed points;
a calculation processing unit for calculating the distance of the orthogonal projection of each point in the point cloud to the plane and comparing the distance with a set threshold Th dist Comparing when the height difference is less than threshold Th dist Considering the point as belonging to the ground point, when the height difference is larger than a threshold Th dist If the point is not the ground point;
the iteration segmentation unit is used for taking all classified ground points as a seed point set of the next iteration and continuing the iteration optimization; and (5) obtaining the iteration times, and finally obtaining the segmented ground point cloud.
As a further limitation of the technical solution of the present invention, the apparatus further includes a plane evaluation module for evaluating the established initial plane model.
As a further limitation of the technical solution of the present invention, the boundary point obtaining module includes a receiving unit, a converting unit, and a dividing obtaining unit;
the receiving unit is used for receiving the segmented ground point cloud;
the conversion unit is used for converting the 3D data point set to an xy plane through coordinate conversion to obtain a plane point set;
the segmentation acquisition unit is used for establishing a minimum bounding box of the data point set; and dividing the ground point cloud by using a rectangular network with set intervals, finding out a boundary grid and extracting boundary points.
As a further limitation of the technical scheme of the invention, the segmentation acquisition unit is specifically used for traversing all points of the plane point set to obtain the maximum value and the minimum value of the X axis and the maximum value and the minimum value of the Y axis and establish a minimum bounding box of the point set; after the ground point cloud is gridded, calculating the row and the column where the point is located according to the coordinate of the current point, adding the point into the grid corresponding to the row and the column, wherein if the number of points in the corresponding grid is zero, the pixel value is 0, and otherwise, the pixel value is 1; judging each grid, if the number of points of the grid is not zero and the number of points in the adjacent grid is not zero, determining the grid as a non-boundary grid, otherwise determining the grid as a boundary grid; and putting the point cloud in the boundary grid into a corresponding storage container to obtain road boundary points.
As a further limitation of the technical scheme of the invention, the boundary line fitting module is used for performing piecewise fitting on the obtained boundary points by using a B-spline curve, filling the boundary points at the vacant positions, and removing part of abnormal points.
In a third aspect, the present invention further provides an intelligent mine card, where the intelligent mine card is provided with a laser radar and a processor, and the processor is capable of executing the method for identifying a travelable area based on laser sensing according to the first aspect.
According to the technical scheme, the invention has the following advantages: (1) The drivable area identification method based on the laser point cloud carries out ground detection based on a ground plane identification and optimization method, identifies and extracts road boundary points by using an equally-spaced rectangular grid division method, fills the boundary points of vacant positions with an obtained boundary point by using a B-spline curve fitting method, simultaneously eliminates the boundary points of abnormal positions, and finally outputs the drivable area of a mine car. The method is simple and reliable, easy to implement and good in recognition effect, and provides an effective method for recognizing the travelable area of the unstructured road in the mountainous area in the current automatic driving field.
(2) The method is based on the identification mode of the laser radar point cloud, has the advantages of being naturally free from the influence of illumination and weather condition changes, is wide in identification range and high in identification precision, and guarantees the feasibility of the algorithm.
In addition, the invention has reliable design principle, simple structure and very wide application prospect.
Therefore, compared with the prior art, the invention has prominent substantive features and remarkable progress, and the beneficial effects of the implementation are also obvious.
Drawings
In order to more clearly illustrate the embodiments or technical solutions in the prior art of the present invention, the drawings used in the description of the embodiments or prior art will be briefly described below, and it is obvious for those skilled in the art that other drawings can be obtained based on these drawings without creative efforts.
FIG. 1 is a schematic flow diagram of a method of one embodiment of the present invention.
FIG. 2 is a diagram of a pseudo code for a ground plane fitting algorithm.
Fig. 3 is a schematic diagram of point cloud meshing.
FIG. 4 is a schematic diagram of a partitioned boundary point cloud grid.
Fig. 5 is a schematic block diagram of an apparatus of one embodiment of the present invention.
Detailed Description
In order to make those skilled in the art better understand the technical solution of the present invention, the technical solution in the embodiment of the present invention will be clearly and completely described below with reference to the drawings in the embodiment of the present invention, and it is obvious that the described embodiment is only a part of the embodiment of the present invention, and not all embodiments. 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.
As shown in fig. 1, an embodiment of the present invention provides a method for identifying a travelable area based on laser sensing, including the following steps:
s1: converting the point cloud from a radar coordinate system to a vehicle body coordinate system and preprocessing point cloud data;
and acquiring front road point cloud data acquired by the laser radar, and performing point cloud downsampling filtering. Traversing all points in the point cloud, eliminating points with coordinate values equal to the NaN value, and eliminating points with coordinate values smaller than a certain distance (considered as the point cloud of the scanned vehicle body). And (4) carrying out coordinate system transformation on the point cloud data, and transferring all point clouds from the radar coordinate system to the vehicle body coordinate system. And (4) defining a plane point cloud to detect the ROI, and intercepting the point cloud positioned in the ROI to obtain a preprocessed point cloud.
S2: performing ground plane fitting by using the preprocessed point cloud, and segmenting ground points from the preprocessed point cloud;
sorting the preprocessed point clouds according to the height of the z axis, removing partial points with abnormal height of the z axis, and marking as non-ground points; points with z-axis height near the ground height are defined as seed points, which are used to build an initial planar model describing the ground.
The concept of Lowest Point Representation (LPR) is introduced. The LPR is the average of the lowest elevation points and ensures that the plane fitting phase is not affected by measurement noise. The LPR is regarded as the lowest point of the whole point cloud P, and the points in the point cloud P with the height within the threshold value are regarded as seed points, and the seed points form a seed point set.
The distance of the orthogonal projection of the point in the point cloud to the plane is less than a threshold Th dist The point is considered to belong to the ground, otherwise it belongs to non-ground. We use a simple linear model for the planar model estimation and then use three singular vectors to describe the spread of the point set in three main directions. Since it is a plane model, a normal vector perpendicular to the plane represents a direction having the smallest variance, and can be found by calculating a singular vector having the smallest singular value. After selecting the seed point set and evaluating the initial plane model, we calculate the distance of the orthogonal projection of each point in the point cloud to the plane, and compare this distance with a set threshold Th dist And comparing, when the height difference is smaller than the threshold value, the point is considered to belong to the ground, and when the height difference is larger than the threshold value, the point is a non-ground point. And all classified ground points are taken as a seed point set of next iteration, and iteration optimization is continued. And finally obtaining the segmented ground point cloud through certain iteration times.
S3: dividing the divided ground points by using a rectangular network with set intervals, finding out a boundary grid and extracting boundary points;
for calculation convenience, a 3D data point set is converted to an xy plane through coordinate conversion to obtain a plane point set. The most common meshing method is uniform meshing, where all points are first traversed to obtain x max ,x min ,y max ,y min And establishing a minimum bounding box of the point set; and putting the data points into the corresponding network units, and establishing a corresponding relation between the data points and the network units. And after the point cloud data is subjected to gridding, corresponding to a pair of binary images with the size of M x N, wherein if the number of points in the corresponding grid is zero, the pixel value is 0, and otherwise, the pixel value is 1. And judging each grid, if the point number of the grid is not zero and the point number in 8 neighborhood grids is not zero, determining the grid as a non-boundary grid, otherwise determining the grid as a boundary grid. And placing the point cloud in the boundary grid into a corresponding storage container.
S4: performing boundary line fitting on the extracted boundary points, filling the boundary points of the vacant positions, and eliminating points with abnormal value deviation larger than a set threshold value;
the road boundary points obtained by the above steps may be discontinuous or even have a relatively large deviation from the abnormal value. And fitting the obtained boundary points by using a B spline curve, filling the boundary points at the vacant positions, and removing part of abnormal points.
S5: and issuing road boundary points identified based on the laser point cloud, and issuing travelable areas of the mine car.
The embodiment of the invention provides a method for identifying a travelable area based on laser perception, which comprises the following steps:
as shown in fig. 2-4, step one: a64-line laser radar is installed right in front of the automatic driving mine car, and in order to more fully acquire the road condition of the environment in front of the short-distance car, the laser radar is installed at an angle inclined downwards by 30 degrees horizontally. And calibrating external parameters from the laser radar to the vehicle body coordinate system after the installation is finished, and recording external parameter data for coordinate transformation of point cloud. The relevant parameters of the laser radar adopted by the invention are shown in the table 1.
TABLE 1
| Parameter(s) | Numerical value |
| Vertical resolution | 64 lines |
| Horizontal resolution | 1024 |
| Field of view | And (2) vertically: 45 ° (+ 22.5 ° to-22.5 °)/horizontal: 360 degree |
| Angular sampling accuracy | And (2) vertically: 0.01 °/level: plus or minus 0.01 degree |
| Frequency of rotation | 10Hz |
Step two: and acquiring front road point cloud data acquired by the laser radar, and performing point cloud down-sampling filtering. Traversing all points in the point cloud, eliminating points with coordinate values equal to the NaN value, and eliminating points with coordinate values smaller than a certain distance (considered as the point cloud of the scanned vehicle body). And (4) carrying out coordinate system transformation on the point cloud data, and transferring all point clouds from the radar coordinate system to the vehicle body coordinate system. And (4) defining a plane point cloud to detect the ROI, and intercepting the point cloud positioned in the ROI to obtain the preprocessed point cloud.
Step three: sorting the preprocessed point cloud according to the height of the z axis, removing partial points with abnormal height of the z axis, and marking as non-ground points; points with z-axis height near the ground height are defined as seed points, which are used to build an initial planar model describing the ground.
The concept of Lowest Point Representation (LPR) is introduced. The LPR is the average of the lowest elevation points and ensures that the plane fitting phase is not affected by measurement noise. The LPR is regarded as the lowest point of the whole point cloud P, and the points in the point cloud P with the height within the threshold value are regarded as seed points, and the seed points form a seed point set.
Step four: next, we need to determine a plane to which the distance of the orthogonal projection of the point in the point cloud is less than a threshold Th dist The point is considered to belong to the ground, otherwise it belongs to the non-ground. We use a simple linear model for the planar model estimation and then use three singular vectors to describe the spread of the point set in three main directions. Since it is a plane model, a normal vector perpendicular to the plane represents a direction having the smallest variance, and can be found by calculating a singular vector having the smallest singular value.
Step five: after selecting the seed point set and evaluating the initial plane model, we calculate the distance of the orthogonal projection of each point in the point cloud to the plane, and compare this distance with a set threshold Th dist And comparing, when the height difference is smaller than the threshold value, the point is considered to belong to the ground, and when the height difference is larger than the threshold value, the point is a non-ground point. And all the classified ground points are taken as a seed point set of the next iteration, and the iteration optimization is continued. And finally obtaining the segmented ground point cloud through certain iteration times.
Step six: and receiving the segmented ground point cloud, firstly establishing a minimum bounding box of the data point set, and segmenting the data point set by using a rectangular network with a given interval.
For calculation convenience, a 3D data point set is converted to an xy plane through coordinate conversion to obtain a plane point set.
As shown in FIG. 3, the most common meshing method is uniform meshing, where all points are first traversed to obtain x max ,x min ,y max ,y min And establishing a minimum bounding box of the point set;
and putting the data points into the corresponding network units, and establishing the corresponding relation between the data points and the network units.
Step seven: after the ground point cloud is meshed, distributing each point to each grid, firstly calculating the row and the column of the point according to the coordinate of the current point, and then adding the point to the grids of the corresponding row and the column.
And after the point cloud data is meshed, corresponding to a binary image with the size of M x N, wherein if the number of points in the corresponding mesh is zero, the pixel value is 0, and otherwise, the pixel value is 1.
And judging each grid, if the point number of the grid is not zero and the point number in 8 neighborhood grids is not zero, determining the grid as a non-boundary grid, otherwise determining the grid as a boundary grid. And putting the point cloud in the boundary grid into a corresponding storage container. As shown in fig. 4.
Step eight: the road boundary points obtained by the above steps may be discontinuous or even have a relatively large deviation of abnormal values. And fitting the obtained boundary points by using a B spline curve, filling the boundary points at the vacant positions, and removing part of abnormal points.
Due to the characteristics that noise is acquired by point cloud and the boundary of the unstructured road in a mining area is not obvious, the boundary points acquired in the steps have certain deviation, and when a cubic B-spline curve is used for carrying out segmentation fitting, no middle control point is set. Except for the starting point P 0 End point P N In addition, the intermediate point will pull the curve towards itself, but will not pass the intermediate point.
When the multi-point cubic B spline is in the segment fitting, P is used 0 、P 1 、P 2 、P 3 Drawing the 1 st sample curve, P 1 、P 2 、P 3 、P 4 Drawing the curve of the No. 2 sample strip, P N-3 、P N-2 、P N-1 、P N And drawing the curve of the (N-2) th sample strip. The curves are connected with each other and have C 2 The stages are sequential.
Step nine: and issuing road boundary points identified based on the laser point cloud, and issuing travelable areas of the mine car.
As shown in fig. 5, an embodiment of the present invention further provides a device for identifying a travelable region based on laser sensing, including a preprocessing module, a ground point segmentation module, a boundary point obtaining module, a boundary line fitting module, and a publishing module;
the preprocessing module is used for converting the point cloud from a radar coordinate system to a vehicle body coordinate system and preprocessing the point cloud data;
the ground point segmentation module is used for performing ground plane fitting by using the preprocessed point cloud and segmenting ground points from the preprocessed point cloud;
the boundary point acquisition module is used for segmenting the segmented ground points by using a rectangular network with set intervals, finding out a boundary grid and extracting boundary points;
the boundary line fitting module is used for performing boundary line fitting on the extracted boundary points, filling the boundary points of the vacant positions and eliminating points with abnormal value deviation larger than a set threshold value;
and the issuing module is used for issuing road boundary points identified based on the laser point cloud and issuing drivable areas of the mine car.
In some embodiments, the preprocessing module includes a sampling filtering unit, a first preprocessing unit, a coordinate transformation unit, and a second preprocessing unit;
the sampling filtering unit is used for acquiring front road point cloud data acquired by the laser radar and performing point cloud downsampling filtering;
the first preprocessing unit is used for traversing all points in the point cloud, eliminating points with coordinate values equal to a NaN value and eliminating points with coordinate values smaller than a first threshold value;
the coordinate conversion unit is used for carrying out coordinate system conversion on the point cloud data and converting the point cloud from the radar coordinate system to the vehicle body coordinate system;
and the second preprocessing unit is used for defining a plane point cloud detection ROI, intercepting the point cloud positioned in the ROI and obtaining a preprocessed point cloud.
In some embodiments, the ground point segmentation module comprises a sorting unit, a generating unit, a plane establishing unit, a calculation processing unit and an iteration segmentation unit;
the sorting unit is used for sorting the preprocessed point cloud according to the height of the z axis;
the generating unit is used for calculating the lowest point of the point cloud, and selecting the point of which the height difference between the Z-axis height and the lowest point in the point cloud is within a second threshold value as a seed point to generate a seed point set;
the plane establishing unit is used for establishing an initial plane model for describing the ground according to the seed points;
a calculation processing unit for calculating the distance of the orthogonal projection of each point in the point cloud to the plane and comparing the distance with a set threshold Th dist Comparing when the height difference is less than threshold Th dist Considering the point as belonging to the ground point, when the height difference is greater than the threshold Th dist If the point is not the ground point;
the iteration segmentation unit is used for taking all classified ground points as a seed point set of the next iteration and continuing the iteration optimization; and (5) obtaining the iteration times, and finally obtaining the segmented ground point cloud.
In some embodiments, the apparatus further comprises a plane evaluation module for evaluating the established initial plane model.
The boundary point acquisition module comprises a receiving unit, a conversion unit and a segmentation acquisition unit;
the receiving unit is used for receiving the segmented ground point cloud;
the conversion unit is used for converting the 3D data point set to an xy plane through coordinate conversion to obtain a plane point set;
the segmentation acquisition unit is used for establishing a minimum bounding box of the data point set; and the ground point cloud is divided by using a rectangular network with set intervals, a boundary grid is found out, and boundary points are extracted.
In some embodiments, the segmentation obtaining unit is specifically configured to traverse all points of the plane point set to obtain a maximum value and a minimum value of an X axis and a maximum value and a minimum value of a Y axis, and establish a minimum bounding box of the point set; after the ground point cloud is meshed, calculating the row and the column of the point according to the coordinate of the current point, adding the point into the grids of corresponding rows and columns, and if the number of points in the corresponding grids is zero, the pixel value is 0, otherwise, the pixel value is 1; judging each grid, if the number of points of the grid is not zero and the number of points in the adjacent grid is not zero, determining the grid as a non-boundary grid, otherwise determining the grid as a boundary grid; and putting the point cloud in the boundary grid into a corresponding storage container to obtain the road boundary points.
In some embodiments, the boundary line fitting module is configured to perform piecewise fitting on the obtained boundary points by using a B-spline, fill up the boundary points at the vacant positions, and remove part of outliers.
The embodiment of the invention also provides an intelligent mine card which is provided with a laser radar and a processor, wherein the processor can execute the identification method of the travelable area based on laser perception.
Although the present invention has been described in detail by referring to the drawings in connection with the preferred embodiments, the present invention is not limited thereto. Various equivalent modifications or substitutions can be made on the embodiments of the present invention by those skilled in the art without departing from the spirit and scope of the present invention, and these modifications or substitutions should be within the scope of the present invention/any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present disclosure and the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the appended claims.
Claims (10)
1. A method for identifying a travelable area based on laser perception is characterized by comprising the following steps:
converting the point cloud from a radar coordinate system to a vehicle body coordinate system and preprocessing point cloud data;
performing ground plane fitting by using the preprocessed point cloud, and segmenting ground points from the preprocessed point cloud;
dividing the divided ground points by using a rectangular network with a set interval, finding out a boundary grid and extracting boundary points;
performing boundary line fitting on the extracted boundary points, filling the boundary points of the vacant positions, and eliminating points with abnormal value deviation larger than a set threshold value;
and issuing road boundary points identified based on the laser point cloud, and issuing travelable areas of the mine car.
2. The method for identifying the travelable region based on laser perception according to claim 1, wherein the step of converting the point cloud from a radar coordinate system to a vehicle body coordinate system and preprocessing the point cloud data comprises:
acquiring front road point cloud data acquired by a laser radar, and performing point cloud down-sampling filtering;
traversing all points in the point cloud, eliminating points with coordinate values equal to the NaN value and eliminating points with coordinate values smaller than a first threshold value;
carrying out coordinate system transformation on the point cloud data, and transferring the point cloud from a radar coordinate system to a vehicle body coordinate system;
and (4) defining a plane point cloud to detect the ROI, and intercepting the point cloud positioned in the ROI to obtain the preprocessed point cloud.
3. The method for identifying the travelable region based on laser perception according to claim 2, wherein the ground plane fitting is performed by using the preprocessed point cloud, and the step of segmenting the ground points from the preprocessed point cloud comprises:
sorting the preprocessed point clouds according to the height of a z axis;
calculating the lowest point of the point cloud, and selecting the point of which the height difference between the Z-axis height and the lowest point in the point cloud is within a second threshold value as a seed point to generate a seed point set;
establishing an initial plane model for describing the ground according to the seed points;
calculating the distance of the orthogonal projection of each point in the point cloud to the plane, and comparing the distance with a set threshold Th dist Comparing when the height difference is less than threshold Th dist Considering the point as belonging to the ground point, when the height difference is greater than the threshold Th dist If the point is not the ground point;
all classified ground points are taken as a seed point set of next iteration, and iteration optimization is continued; and (5) obtaining the iteration times, and finally obtaining the segmented ground point cloud.
4. The method for identifying a drivable area based on laser perception as claimed in claim 3, characterized in that the step of building an initial plane model describing the ground from the seed points is followed by:
and evaluating the established initial plane model.
5. The method for identifying a travelable region based on laser sensing of claim 4, wherein the step of dividing the divided ground points using a rectangular network of set intervals, finding a boundary mesh and extracting boundary points comprises:
receiving the segmented ground point cloud;
converting the 3D data point set to an xy plane through coordinate conversion to obtain a plane point set;
establishing a minimum bounding box of the data point set; and the ground point cloud is divided by using a rectangular network with set intervals, a boundary grid is found out, and boundary points are extracted.
6. The identification method of the travelable region based on laser perception according to claim 5, characterized by establishing a minimum bounding box of the data point set; the method comprises the following steps of dividing ground point cloud by using a rectangular network with set intervals, finding out boundary grids and extracting boundary points:
traversing all points of the plane point set to obtain the maximum value and the minimum value of an X axis and the maximum value and the minimum value of a Y axis and establishing a minimum bounding box of the point set;
and putting the data points into corresponding network units, finding out boundary grids and extracting boundary points.
7. The method for identifying a travelable area based on laser perception according to claim 6, wherein the step of placing data points into corresponding network elements, finding a boundary grid and extracting boundary points comprises:
after the ground point cloud is gridded, calculating the row and the column where the point is located according to the coordinate of the current point, adding the point into the grid corresponding to the row and the column, wherein if the number of points in the corresponding grid is zero, the pixel value is 0, and otherwise, the pixel value is 1;
judging each grid, if the number of points of the grid is not zero and the number of points in the adjacent grid is not zero, determining the grid as a non-boundary grid, otherwise determining the grid as a boundary grid;
and putting the point cloud in the boundary grid into a corresponding storage container to obtain road boundary points.
8. The method for identifying the travelable region based on laser perception according to claim 7, wherein the step of performing boundary line fitting on the extracted boundary points, filling the boundary points at the vacant positions, and eliminating the points with the abnormal value deviation larger than the set threshold value comprises:
and performing segmentation fitting on the obtained boundary points by using a B spline curve, filling the boundary points at the vacant positions, and removing part of abnormal points.
9. A travelable region identification device based on laser perception is characterized by comprising a preprocessing module, a ground point segmentation module, a boundary point acquisition module, a boundary line fitting module and a publishing module;
the preprocessing module is used for converting the point cloud from a radar coordinate system to a vehicle body coordinate system and preprocessing the point cloud data;
the ground point segmentation module is used for performing ground plane fitting by using the preprocessed point cloud and segmenting ground points from the preprocessed point cloud;
the boundary point acquisition module is used for segmenting the segmented ground points by using a rectangular network with set intervals, finding out a boundary grid and extracting boundary points;
the boundary line fitting module is used for performing boundary line fitting on the extracted boundary points, filling the boundary points of the vacant positions and eliminating points with abnormal value deviation larger than a set threshold value;
and the issuing module is used for issuing road boundary points identified based on the laser point cloud and issuing travelable areas of the mine car.
10. A smart mine card provided with a lidar, wherein the smart mine card comprises a processor capable of executing the method for identifying a travelable region based on laser perception according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211531348.4A CN115761682A (en) | 2022-12-01 | 2022-12-01 | Method and device for identifying travelable area based on laser perception and intelligent mine card |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211531348.4A CN115761682A (en) | 2022-12-01 | 2022-12-01 | Method and device for identifying travelable area based on laser perception and intelligent mine card |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115761682A true CN115761682A (en) | 2023-03-07 |
Family
ID=85342425
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211531348.4A Pending CN115761682A (en) | 2022-12-01 | 2022-12-01 | Method and device for identifying travelable area based on laser perception and intelligent mine card |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115761682A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116612451A (en) * | 2023-07-20 | 2023-08-18 | 城市之光(深圳)无人驾驶有限公司 | Road edge identification method, device and equipment for unmanned sweeper and storage medium |
| CN117392632A (en) * | 2023-12-11 | 2024-01-12 | 中交第二公路勘察设计研究院有限公司 | Road element change monitoring method and device |
| CN117437214A (en) * | 2023-11-25 | 2024-01-23 | 兰州交通大学 | Rail surface extraction and foreign matter identification method based on bidirectional cloth simulation point cloud |
-
2022
- 2022-12-01 CN CN202211531348.4A patent/CN115761682A/en active Pending
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116612451A (en) * | 2023-07-20 | 2023-08-18 | 城市之光(深圳)无人驾驶有限公司 | Road edge identification method, device and equipment for unmanned sweeper and storage medium |
| CN116612451B (en) * | 2023-07-20 | 2023-09-29 | 城市之光(深圳)无人驾驶有限公司 | Road edge identification method, device and equipment for unmanned sweeper and storage medium |
| CN117437214A (en) * | 2023-11-25 | 2024-01-23 | 兰州交通大学 | Rail surface extraction and foreign matter identification method based on bidirectional cloth simulation point cloud |
| CN117392632A (en) * | 2023-12-11 | 2024-01-12 | 中交第二公路勘察设计研究院有限公司 | Road element change monitoring method and device |
| CN117392632B (en) * | 2023-12-11 | 2024-03-15 | 中交第二公路勘察设计研究院有限公司 | Road element change monitoring method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115761682A (en) | Method and device for identifying travelable area based on laser perception and intelligent mine card | |
| CN110320504B (en) | Unstructured road detection method based on laser radar point cloud statistical geometric model | |
| CN105404844B (en) | A kind of Method for Road Boundary Detection based on multi-line laser radar | |
| CN104392212B (en) | The road information detection and front vehicles recognition methods of a kind of view-based access control model | |
| CN108873013B (en) | Method for acquiring passable road area by adopting multi-line laser radar | |
| CN112380312B (en) | Laser map updating method based on grid detection, terminal and computer equipment | |
| CN102768022B (en) | Tunnel surrounding rock deformation detection method adopting digital camera technique | |
| CN114842450B (en) | Method, device and equipment for detecting drivable area | |
| CN103226833A (en) | Point cloud data partitioning method based on three-dimensional laser radar | |
| CN111325138B (en) | Road boundary real-time detection method based on point cloud local concave-convex characteristics | |
| CN111354083B (en) | Progressive building extraction method based on original laser point cloud | |
| CN112731440A (en) | High-speed railway slope deformation detection method and device | |
| CN110488151B (en) | Remote sensing technology-based power transmission line vegetation hazard early warning system and method | |
| CN116071729A (en) | Method and device for detecting drivable area and road edge and related equipment | |
| CN112014856A (en) | Road edge extraction method and device suitable for cross road section | |
| CN112926468A (en) | Tidal flat elevation automatic extraction method | |
| CN113345094A (en) | Electric power corridor safety distance analysis method and system based on three-dimensional point cloud | |
| CN113379919A (en) | Vegetation canopy height rapid extraction method based on unmanned aerial vehicle RGB camera | |
| CN116027339A (en) | Mine underground map building and positioning method based on laser radar | |
| CN115657049A (en) | Tunnel vehicle-mounted laser radar positioning and deviation rectifying method and system | |
| CN117932333A (en) | Urban building height extraction method considering different terrain scenes | |
| CN114092658A (en) | High-precision map construction method | |
| CN113176546B (en) | Method for filtering background of road side radar in real time based on NURBS modeling | |
| CN117634684A (en) | Urban rail transit environment vibration influence prediction method and system based on GIS spatial analysis | |
| CN113420670A (en) | Environment-friendly supervision method for changing power transmission and transformation line migration based on high-resolution remote sensing |
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 |