CN115236628A - A method for detecting residual cargo in carriages based on lidar - Google Patents

Abstract

The invention relates to a method for detecting carriage residual cargos based on a laser radar, which comprises the steps of extracting carriage data in single-frame original data; extracting a carriage target point cloud, and performing inclination correction on the acquired single-frame point cloud outline; acquiring the current vehicle speed, converting a coordinate system taking a radar as an origin into a coordinate system taking the bottom of the carriage as a center, and splicing the single-frame point clouds to form a carriage integral point cloud picture; denoising and simplifying the carriage point cloud data; smoothing the spliced compartment point cloud data; slicing the carriage point cloud data; projecting the point cloud between two adjacent slices, and extracting a section outline; for the extracted section outline, the section area is obtained, and the section area is multiplied by the slice interval to obtain the point cloud volume of the section; and accumulating and calculating the volume of the residual cargos in the carriage. The automatic detection device can automatically detect the residual condition of the goods in the train carriage without stopping the train, does not need workers to climb into the carriage for observation, reduces the labor intensity of the workers, improves the safety of the work, and has high automatic detection efficiency and good accuracy.

Description

A method for detecting residual cargo in carriages based on lidar

technical field

The invention relates to a method for detecting residual cargo in a carriage, in particular to a method for detecting residual cargo in a carriage based on laser radar.

Background technique

Cargo residual cargo is a common phenomenon for railway freight vehicles. Due to the shape and characteristics of the objects loaded in the car, or due to weather and other reasons, it is impossible to completely unload the cargo in the car during unloading operations. For residual goods, workers are often required to climb to the top of the carriage with the help of tools to observe the residual condition of the goods in the box, and then arrange a corresponding number of workers to enter the carriage for cleaning according to the amount of residue to avoid loss of tons. Such manual detection methods not only have potential safety hazards and affect the normal operation of trains, but also increase the labor intensity of workers and reduce work efficiency.

SUMMARY OF THE INVENTION

Aiming at the above problems in the prior art, the present invention provides a method for detecting residual cargo in a carriage based on lidar, which does not affect train operation, is convenient for detecting residuals, does not require workers to board the train for inspection, and has high safety and efficiency.

In order to achieve the above objects, the present invention provides the following technical solutions: a method for detecting residual cargo in a carriage based on lidar, and a method for detecting residual cargo in a carriage based on lidar, which is characterized by comprising the following steps:

S10, using the Euclidean distance-based cluster analysis method to extract the compartment data in the single frame of original data, and using the three-dimensional coordinate relationship to determine the front and rear boundaries of the compartment, and obtain the data information of the first frame and the last frame;

S20, using the method of conditional filtering to extract the target point cloud of the carriage, and performing tilt correction on the collected single-frame point cloud contour;

S30, use the preset radar to obtain the current vehicle speed, convert the coordinate system with the radar as the origin to the coordinate system with the bottom of the carriage as the center, and use the displacement fusion algorithm to stitch the point clouds of a single frame to form the overall point cloud image of the carriage;

S40, using statistical filtering and voxelized grid method to denoise and simplify the point cloud data of the carriage;

S50, using the moving least squares method to smooth the spliced point cloud data of the carriage;

S60, slicing the point cloud data of the carriage along the running direction of the vehicle;

S70, project the point cloud between two adjacent slices, and use the alpha algorithm and the ray 360-degree algorithm to extract the cross-sectional contour;

S80 , for the extracted cross-sectional profile, use the shoelace theorem to obtain the cross-sectional area, and multiply it by the slice spacing to obtain the volume of the segment to the point cloud; accumulate and calculate the volume of residual cargo in the carriage.

Further, performing conditional filtering and contour correction on the point cloud in the step S20 includes: according to the frame data obtained by scanning, selecting the static point cloud scanned in the frame, looking for the same column of point cloud data, analyzing its coordinate values, and fitting calculation. The deflection angle is used to correct the rotation of the overall point cloud data.

Further, in the step S30, the coordinate system transformation and point cloud splicing include: transforming the coordinate system with the radar as the origin into the coordinate system with the carriage as the origin, determining the coordinate value transformation relationship according to the radar installation position relationship, and completing the coordinate system. System conversion; the point cloud splicing method based on mobile displacement fusion, by assigning a new Z-axis coordinate value to the point cloud, the point cloud data of all frames are uniformly corrected to the same frame coordinate system to complete the point cloud splicing.

Further, the method for statistical filtering and voxelization grid in the step S40 includes: the average spacing distribution relationship from the sampling point to the neighborhood point conforms to the Gaussian distribution function, and the outlier point cloud data is sorted by setting a reasonable threshold. Elimination: downsample the point cloud data using the voxel grid method, and drop the point cloud data into a grid of a specified size, each grid only retains the point cloud closest to the center of the grid, and the rest of the grid The point cloud is deleted, so as to achieve the effect of simplifying the point cloud.

Further, slicing the point cloud data in the step S60 includes: intercepting the point cloud data of the carriage with a tangent parallel to the cross section of the carriage at a certain interval along the running direction of the vehicle to form a segment of scattered point cloud data. . The larger the spacing, the more point cloud segments will be cut, and vice versa.

Further, extracting the point cloud contour in the step S70 includes: projecting the point cloud between adjacent slices, and using an alpha algorithm to determine a rough boundary contour; then using a ray 360-degree scanning algorithm to perform secondary discrimination on the boundary contour, and eliminating the Unqualified boundary point cloud data. Improve the accuracy of boundary contours.

Further, calculating the point cloud volume in the step S80 includes: using the shoelace theorem to obtain the point cloud cross-sectional area of the irregular polygon, multiplying it by the slice interval to obtain the volume of the point cloud data, and obtaining the overall point cloud volume of the carriage by accumulation. ; Use the difference between the point cloud volume of the empty car and the volume of the car when there is residual cargo to obtain the volume of the residual cargo in the car.

Compared with the prior art, the present invention can automatically detect the residual condition of goods in the train compartment without stopping the train, without the need for workers to climb into the compartment to observe, reduce the labor intensity of workers and improve the safety of work, the automatic detection efficiency is high and the The accuracy is good; the present invention can not only identify the existing area and volume of the residual cargo in the railway carriage, but also can detect the materials such as loose sand and loose rocks loaded on the road open-top truck. quality to determine the overloading of the truck.

Description of drawings

Fig. 1 is the flow chart of the present invention;

FIG. 2 is a data diagram of a carriage according to an embodiment of the present invention;

FIG. 3 is a diagram for judging a compartment boundary according to an embodiment of the present invention;

FIG. 4 is a diagram showing the reason for the inclination of the point cloud contour of the carriage according to the embodiment of the present invention;

5 is a correction diagram of a single frame of point cloud data according to an embodiment of the present invention;

6 is a schematic diagram of coordinate system conversion according to an embodiment of the present invention;

FIG. 7 is an embodiment of the present invention splicing point cloud images of original carriages based on moving displacement;

8 is a result diagram of a voxelized grid method and point cloud smoothing according to an embodiment of the present invention;

FIG. 9 is a sectional view of the outline of an empty carriage according to an embodiment of the present invention;

FIG. 10 is a cross-sectional view of the outline of the carriage when there is residue in the embodiment of the present invention;

FIG. 11 is a schematic diagram of an area calculation principle of an irregular polygon according to an embodiment of the present invention.

Detailed ways

The present invention will be further described below in conjunction with the accompanying drawings.

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

As shown in Figure 1, the present invention provides a technical solution, a method for detecting residual cargo in a carriage based on lidar:

S10, using the Euclidean distance-based cluster analysis method to extract the compartment data in the single frame of original data, and using the three-dimensional coordinate relationship to determine the front and rear boundaries of the compartment, and obtain the data information of the first frame and the last frame;

The cluster analysis algorithm based on Euclidean distance is used to preprocess the initial data of the point cloud. The clustering conditions are as follows: the search radius of the nearest neighbor search is 0.2 meters, the minimum number of clustering points is 1000, and the maximum number of clustering points is 5000; the specific implementation steps as follows:

(1) Find a point P ₁₀ in the space, find the n points closest to him in kdTree, and judge the distance from these n points to P ₁₀ . Put points P ₁₂ , P ₁₃ , P ₁₄ . . . whose distances are less than the threshold r in class Q;

(2) Find a point P ₁₂ in Q(P ₁₀ ), repeat (1);

(3) Find a point in Q (P ₁₀ , P ₁₂ ), repeat (1), find P ₂₂ , P ₂₃ , P ₂₄ ...... all put into Q;

(4) When Q can no longer have new points added, the search is completed.

As shown in Figure 2 and Figure 3, if the XOY plane of the radar is completely perpendicular to the running direction of the train, assuming that the width of a carriage is 3.3 meters and the height is 2.8 meters, in the judgment of the carriage boundary, point M is the highest point of the carriage, and the coordinates of point M are is (x _m , y _max , z _m ), point N is the lowest point of the carriage, and the coordinates of point N are (X _n , y _min , z _n ), according to the three-dimensional coordinate relationship:

y _max -y _min =2.8,

Among them, y _max is the y coordinate of the highest point M, y _min is the y coordinate of the lowest point N, and the height of the carriage is 2.8 meters. After processing the point cloud data of the carriage, hide the left and right side walls of the carriage by 0.5 meters. If the change range of y _max and y _min of the point cloud data is small, that is, the difference between y _max and y _min is greater than a certain threshold t, then it is judged that the current compartment data is the compartment boundary data, otherwise it is judged that it is not the compartment boundary data. When the compartment boundary information appears for the first time, it starts to record data information; when the compartment data appears for the last time, it stops recording data information and obtains a complete compartment data information.

S20, using the method of conditional filtering to extract the target point cloud of the carriage, and performing contour inclination correction on the collected point cloud data of a single frame;

If the XOY plane of the radar is completely perpendicular to the running direction of the train, the laser points scanned in the same row should be on the same horizontal line, as shown in Figure 4(a); if the XOY plane of the radar is not perpendicular to the running direction of the train, there is an inclined angle, then The points of the same column of data also show a tilted state, as shown in Figure 4(b); the method based on static point cloud correction in the present invention, the steps are as follows:

(1) arbitrarily take a frame of point cloud data of the measurement scan, select the stationary object around the scanned object as the reference frame (the wall next to the track is selected by the present invention), and extract the overall data of the reference frame point cloud;

(2) Inverse operation is performed according to the three-dimensional coordinates of the point cloud data to obtain the number of columns in the radar scan of the point cloud;

(3) Extract the data Q={P ₀ , P ₁ ...... P ₁₅ } in the same column, the X and Y coordinate values of these data are approximately the same, and the Z-axis data is gradually changed according to a certain trend;

(4) carry out fitting line according to the scatter image of Z-axis coordinate, calculate and fit line equation, utilize atan to obtain the included angle θ _z with respect to Z-axis;

(5) Determine the rotation matrix and the rotation direction through the angle θ, and realize the rotation transformation of the frame point cloud.

As shown in Figure 5, the white point cloud is the data after correction, and the gray point cloud is the data before correction.

S30, use the preset radar to obtain the current vehicle speed, convert the coordinate system with the radar as the origin to the coordinate system with the bottom of the car as the center, and use the displacement fusion algorithm to splicing the point clouds of a single frame to form an overall point cloud image of the car;

As shown in Figure 6, point A is a scanning point in the car, h is the vertical distance between the lidar and the bottom of the car, in the radar coordinate system, the coordinates of point A are processed by the lidar and the output coordinates are (x, y, z) ), after the coordinate system is converted, take O ₁ as the coordinate origin, OO ₁ is on the same straight line, and the coordinates of point A are (x ₁ , y ₁ , z ₁ ), according to the positional geometric relationship,

As shown in FIG. 7 , the present application assigns a new Z value to the Z-axis coordinate of the scanned point cloud based on the method of moving displacement fusion compensation. According to the total number of scanned point cloud frames, each frame of point cloud data is translated in the direction of the train according to the scanning order, and the complete point cloud image based on the carriage coordinate system is obtained. Assuming that the speed of the train moving at a constant speed is V, the operating frequency of the lidar is 10Hz, and the distance the train travels along the Z axis in 0.1 second is V/10, and the displacement compensation is performed on the point cloud data:

Z _ij ′=Z _ij ±(ni)*v/10

Among them, Z _ij ′ represents the Z coordinate value of the jth point in the point cloud data of the ith frame after displacement compensation, and Z _i represents the Z of the jth point in the point cloud data of the ith frame of the ith frame of data in the original coordinate system. Coordinate value, n is the number of recorded point cloud frames, i is the current frame number. When the Z-axis direction of the lidar is the same as the running direction of the train, the above formula takes "+", otherwise, it takes "-".

S40, using statistical filtering and voxelized grid method to denoise and simplify the point cloud data of the carriage;

As shown in Figure 8, the statistical filtering steps are as follows:

(1) Establish point cloud topological structure relationship for target point cloud through KD-tree tree;

(2) Find the neighborhood of the sampling point through the index, and calculate the average Euclidean distance from each sampling point to the neighborhood range point;

(3) Calculate the average distance from all points in the point cloud dataset to the neighborhood;

(4) The obtained distance distribution conforms to the Gaussian function, so as to calculate the mean μ and the standard deviation σ;

(5) Set the threshold to eliminate noise data. According to the distribution characteristics of the Gaussian function, the points in the range of (μ-σ*std, μ+σ*std) belong to the valid point cloud, where std is called the standard deviation multiple, which is used to adjust the set threshold range. If the value of a sampling point d _i is greater than μ+σ*std or less than μ-σ*std, it is determined as a noise point and filtered.

The voxelized grid selects a grid with a size of 0.1m to simplify the point cloud data.

S50, using the moving least squares method to smooth the spliced carriage point cloud data; the basic steps of the moving least squares method are as follows:

(1) Input the fitted data point area, and perform grid operation on the input area;

(2) Determine the range of the influence area of the grid point x, and determine the number of nodes affected within this range; calculate the node value of the grid point according to the formula;

(4) Traverse all grid points, and repeat steps (2) and (3);

(5) Connect the calculated values of grid points to form a smooth curve.

S60, slicing the point cloud data of the carriage along the running direction of the vehicle;

Since the forward direction of the train is the positive half-axis of the Z-axis, the Z-axis is selected as the cutting direction, and the point cloud of the car model is divided into several segments. The density of the point cloud is limited, and it is difficult to form the shape outline of the point cloud simply by relying on a thin plane, so this plane needs to have a certain "thickness". An equidistant point cloud segment is formed between two adjacent slices, so the outline of the point cloud can be obtained by directly projecting the point cloud between the two slice surfaces:

Among them, i is the serial number of the slice, Z _i is the position of the tangent plane, l is the interval distance between the slices, Z _min and Z _max represent the minimum and maximum values of the point cloud coordinates in the Z-axis direction, respectively, and n is the number of tangent planes. number.

S70, project the point cloud between two adjacent slices, and use the alpha algorithm and the ray 360-degree algorithm to extract the cross-sectional contour;

Project the generated point cloud on the XOY plane, use the alpha algorithm to extract the rough outline of the point cloud, and then filter the outline through the ray 360 degree algorithm, as shown in Figure 9 and Figure 10, the ray 360 degree algorithm steps are as follows:

(1) Calculate the center of gravity O of the required point cloud

(2) Take any point P ₀ (x ₀ , y ₀ ) as the initial scanning point, connect OP ₀ as the reference scanning ray, and scan the remaining data points in the counterclockwise direction of the scanning line;

(3) Calculate the angle θ _i between the remaining scanning points and the reference line;

(4) Sort according to the size of the required angle θ _i , and then connect the point clouds in order to form the outline of the point cloud.

S80, for the extracted cross-sectional profile, use the shoelace theorem to obtain the cross-sectional area, and multiply it by the slice spacing to obtain the volume of the segment to the point cloud; through accumulation, calculate the volume of the residual cargo in the carriage.

As shown in Figure 11, the shoelace theorem is to calculate the area of the closed image by the determinant when the vertex coordinates of the polygon are known. It is stipulated that the irregular section is composed of n vertices P ₁ , P ₂ ... P _n . According to the ray scanning method, the sorted points are connected end to end in a counterclockwise order to form a closed polygon, which is denoted as P _i , which is composed of the above The derivation process shows that the calculated cross-sectional area A _i is:

Among them, A _i is the area of the i-th polygonal section, the coordinates of the polygon vertex P _i are (x _i , y _i ), x _n =x ₁ , y _n =y ₁ .

It will be apparent to those skilled in the art that the present invention is not limited to the details of the above-described exemplary embodiments, but that the present invention may be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments are to be regarded in all respects as illustrative and not restrictive, and the scope of the invention is to be defined by the appended claims rather than the foregoing description, which are therefore intended to fall within the scope of the claims. All changes within the meaning and scope of the equivalents of , are included in the present invention. Any reference signs in the claims shall not be construed as limiting the involved claim.

The above are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any minor modifications, equivalent replacements and improvements made to the above embodiments according to the technical essence of the present invention shall be included in the technology of the present invention. within the scope of the program.

Claims (7)

1. A method for detecting residual cargo in a carriage based on lidar, comprising the following steps: S10, using the Euclidean distance-based cluster analysis method to extract the compartment data in the single frame of original data, and using the three-dimensional coordinate relationship to determine the front and rear boundaries of the compartment, and obtain the data information of the first frame and the last frame; S20, using the method of conditional filtering to extract the target point cloud of the carriage, and performing tilt correction on the collected single-frame point cloud contour; S30, use the preset radar to obtain the current vehicle speed, convert the coordinate system with the radar as the origin to the coordinate system with the bottom of the carriage as the center, and use the displacement fusion algorithm to stitch the point clouds of a single frame to form the overall point cloud image of the carriage; S40, using statistical filtering and voxelized grid method to denoise and simplify the point cloud data of the carriage; S50, using the moving least squares method to smooth the spliced point cloud data of the carriage; S60, slicing the point cloud data of the carriage along the running direction of the vehicle; S70, project the point cloud between two adjacent slices, and use the alpha algorithm and the ray 360-degree algorithm to extract the cross-sectional contour; S80 , for the extracted cross-sectional profile, use the shoelace theorem to obtain the cross-sectional area, and multiply it by the slice spacing to obtain the volume of the segment to the point cloud; accumulate and calculate the volume of residual cargo in the carriage. 2 . The method for detecting residual cargo in a carriage based on lidar according to claim 1 , wherein, performing condition filtering and contour correction on the point cloud in the step S20 comprises: selecting the frame data obtained by scanning according to the frame data obtained by scanning. For the static point cloud scanned by the frame, find the same column of point cloud data, analyze its coordinate values, fit and calculate the deflection angle, and perform rotation correction on the overall point cloud data. 3. A method for detecting residual cargo in a carriage based on lidar according to claim 1, wherein the step S30 for transforming the coordinate system and splicing the point cloud comprises: transforming the coordinate system based on the radar as the origin into The coordinate system with the carriage as the origin, according to the radar installation position relationship, determine the coordinate value transformation relationship, and complete the coordinate system transformation; the point cloud splicing method based on mobile displacement fusion, by assigning a new Z-axis coordinate value to the point cloud, all frame points The cloud data is uniformly corrected to the same frame coordinate system to complete the point cloud stitching. 4 . The method for detecting residual cargo in a carriage based on lidar according to claim 1 , wherein the method for statistical filtering and voxelization in the step S40 comprises: averaging the sampling point to the neighborhood point. 5 . The distance distribution relationship conforms to the Gaussian distribution function. By setting the threshold, the outlier point cloud data is eliminated; the voxel grid method is used to downsample the point cloud data, and the point cloud data falls into the grid of the specified size. Only the point cloud closest to the center of the grid is retained for each grid, and the rest of the point clouds in the grid are deleted. 5 . The method for detecting residual cargo in a carriage based on lidar according to claim 1 , wherein slicing the point cloud data in the step S60 comprises: using spaced distances along the running direction of the vehicle that are parallel to the cross-section of the carriage. 6 . The section intercepts the point cloud data of the carriage to form sections of scattered point cloud data. 6 . The method for detecting residual cargo in a carriage based on lidar according to claim 1 , wherein extracting the point cloud contour in the step S70 comprises: projecting the point cloud between adjacent slices, and using an alpha algorithm. 7 . Determine the rough boundary contour; then use the ray 360-degree scanning algorithm to perform secondary discrimination on the boundary contour, and eliminate the boundary point cloud data that does not meet the conditions. 7 . The method for detecting residual cargo in a carriage based on lidar according to claim 1 , wherein calculating the point cloud volume in the step S80 comprises: using the Shoelace theorem to obtain the point cloud cross-sectional area of an irregular polygon, 7 . The volume of the point cloud data is obtained by multiplying it by the slice interval, and the overall point cloud volume of the carriage is obtained by accumulation; the difference between the point cloud volume of the empty carriage and the volume of the carriage when there is residual cargo is used to obtain the volume of the residual cargo in the carriage.

Images