A method for positioning a truck bed rail based on a laser scanner

Abstract

The application discloses a truck compartment tie positioning method based on a laser scanner, solves the problem of tie positioning of a rear-opening compartment truck in an automatic loading and unloading vehicle perception module in an intelligent loading and unloading robot scene, obtains tie position information by obtaining point cloud data of a to-be-measured compartment, running a first positioning algorithm on two-dimensional point cloud data and running a second positioning algorithm on three-dimensional point cloud data, is not limited to two-dimensional point cloud data or three-dimensional point cloud data of a vehicle, saves sensor cost, has fast algorithm operation speed, does not need to call a complicated algorithm, can be used in a small-resource scene such as an embedded scene, and the tailing effect of point cloud is processed to prevent the tailing effect from affecting tie positioning.

Description

Technical Field

[0001] This invention relates to the field of vehicle inspection, and more particularly to a method for locating the tie rods of a truck body based on a laser scanner. Background Technology

[0002] In the intelligent loading and unloading robot market, loading and unloading is a pain point in the industry. With the global push for digital and intelligent transformation and upgrading, automated loading and unloading vehicles are gradually becoming a new development direction. In intelligent loading and unloading scenarios, the perception system is a crucial component. It models the vehicle using vision, ultrasonic waves, or LiDAR positioning, locates the cargo compartment space, identifies the cargo compartment coordinates and tie rod coordinates, and then operates the robot's grippers or unloading port to enter the cargo compartment for operation. The positioning of the tie rods is a critical component, affecting the safety of the entire loading and unloading process.

[0003] Traditional loading and unloading methods for vehicles with reinforcing ribs are mainly manual, with humans controlling robots to enter the cargo compartment or guiding the vehicle forward and backward for loading operations. This method is labor-intensive, costly, and easily influenced by subjective human factors.

[0004] The visual method for locating the car body's reinforcing bars has limited camera field of view and blind spots, requiring fusion of multiple cameras and is prone to calibration errors. The cameras are also affected by ambient light and cannot operate in low-light conditions. Furthermore, the reinforcing bars are often similar in color to the car body and are mostly between 1-5cm in length, making them difficult to distinguish.

[0005] Using 3D multi-line radar or a gimbal-driven laser scanner to acquire 3D point clouds typically results in a massive amount of point cloud data, complex algorithms, and long computation times, making it unsuitable for scenarios with limited computing resources. Furthermore, the laser trailing effect can interfere with the point cloud data from the tie rods and the vehicle body, affecting positioning accuracy.

[0006] Therefore, finding a method for quickly and accurately locating the car body bracing has become an urgent technical problem to be solved. Summary of the Invention

[0007] The purpose of this invention is to provide a method for locating the tie rods of a truck body based on a laser scanner, thereby solving the above-mentioned problems.

[0008] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0009] This invention discloses a method for locating the tie rods of a truck body based on a laser scanner, comprising the following steps:

[0010] (1) In a rectangular testing area, a support is set up, and a first laser scanner and a pan-tilt unit are installed on the support, and a second laser scanner is installed on the pan-tilt unit;

[0011] (2) After the vehicle to be tested enters the detection area, it stops so that the first laser scanner and the second laser scanner are located on the top of the vehicle to be tested.

[0012] (3) Scan the vehicle using a first laser scanner or a second laser scanner;

[0013] (4) When the vehicle is parked at a straight angle, the first laser scanner scans the vehicle's length and height contours in a two-dimensional scan. The origin of the coordinate system is the ground directly below the first laser scanner, the X-axis points towards the rear of the vehicle, and the Y-axis points directly upward. Then, the first positioning algorithm is used, the steps of which include:

[0014] 1) Determine the height of the bottom of the truck bed from the ground;

[0015] 2) Detect the position of the vehicle's point cloud truck bed;

[0016] 3) Determine the position of the trailer's tie rod and obtain point cloud data;

[0017] 4) The data of the point cloud step-by-step flag array is segmented. The adjacent distance of the X coordinate is less than 20cm and it is recorded as a segment. The start position and end position of the segment are recorded. The number of segments equals the number of tendons. The position of the segment is the position of the tendon. In the case of no tail, the difference between the start and end of the segment is the width of the tendon.

[0018] 5) The point cloud caused by the trailing only affects the ranging at the far end of the laser beam, while the ranging at the near end of the laser beam is accurate. Divide the point cloud of the truck bed into positive and negative half axes along the X-axis. The positive half axis data is based on the side with the smaller X-axis data in the segment, and the negative half axis data is based on the side with the larger X-axis data in the segment. The distance between the reference value and the maximum value of the Y-axis in the segment is multiplied by 2 to obtain the width of the beam. The width is limited to 5cm. If the beam width is greater than 5cm, then 5cm is taken as the beam width. Finally, the position and width range of the beam on the X-axis are obtained by calculation. The identified beam position is marked in the point cloud image to distinguish it from the front and rear baffles, providing obstacle avoidance positioning for the unloading port or robot gripper.

[0019] When the vehicle is parked at an angle, and the default maximum deflection angle is 30° (this default maximum deflection angle can be changed according to the actual situation), the second laser scan is performed by the gimbal, performing a three-dimensional scan. The origin of the coordinate system is the ground directly below the laser radar, the Y-axis points towards the rear of the vehicle, the X-axis points to the left side of the vehicle, and the Z-axis points directly upward. Then, the second positioning algorithm is used, the steps of which include:

[0020] 1) Determine the vehicle's deflection angle;

[0021] 2) Vehicle point cloud data correction: The vehicle point cloud data is transformed according to the deflection angle.

[0022] 3) Remove the point cloud data from both side panels;

[0023] 4) Project the point cloud data from the vehicle onto the YZ coordinate system, that is, assign a value of 0 to the X-axis of the point cloud data;

[0024] 5) Determine the height of the bottom of the truck bed from the ground;

[0025] 6) Detect the position of the vehicle's point cloud truck bed;

[0026] 7) Determine the location of the tie rods in the truck bed;

[0027] 8) The point cloud step-by-step flag array data is segmented. The adjacent distance of the Y coordinate is less than 20cm is recorded as a segment. The start position and end position of the segment are recorded. The number of segments is the number of tendons. The position of the segment is the position of the tendon. In the case of no tail, the difference between the start and end of the segment is the width of the tendon.

[0028] 9) Addressing the impact of trailing on the width of the laser beam: The laser beam has a certain divergence angle, and the spot size increases with distance. When there are two objects in front and behind, and the laser spot hits the edge of the object in front, part of the laser spot may hit the object behind. Therefore, a distance between the two is given.

[0029] 10) Process segmented data. The point cloud caused by the trailing effect only affects the ranging at the laser-far end of the tie rod, while the ranging at the laser-near end remains accurate. Divide the truck bed point cloud along the Y-axis into positive and negative halves. The positive half-axis data is based on the smaller Y-axis side within the segment, and the negative half-axis data is based on the larger Y-axis side within the segment. Multiply the distance of the Y-coordinate between the reference and the maximum Z-axis value within the segment by 2 to obtain the tie rod width. A width limit of 5cm is used; if the tie rod width is greater than 5cm, 5cm is taken as the tie rod width. Finally, the distance and width range of the tie rod on the Y-axis are calculated. The identified tie rod position is then marked on the point cloud image, distinguishing it from the front and rear baffles, to provide obstacle avoidance positioning for the unloading port or robot gripper.

[0030] Furthermore, when the first positioning algorithm determines the height of the bottom of the truck bed from the ground, it takes the point cloud data of the X-axis range (Xmax-1000, Xmax) within the detection range, that is, the point cloud data of 1 meter at the rear of the truck bed, and takes the data point Ymin with the smallest Y value as the height of the bottom of the truck bed from the ground. The data unit is mm if there is no label.

[0031] Furthermore, when the first positioning algorithm detects the position of the vehicle's point cloud truck bed, it searches along the negative direction of the X-axis from Xmax to find data points (x1, y1) with Y coordinates greater than Ymin+300, and there are no data points less than Ymin+300 in the range (x1-400, x1-150). These data points are the positions of the front baffle and the point cloud data between the front baffle position and the distance from Xmax.

[0032] Furthermore, the first positioning algorithm determines the position of the truck bed tie rod, takes data in the X-axis range (front baffle +300, rear baffle -300) and the Y-axis range of truck bed bottom height above 300, and establishes a point cloud distribution flag array with an array size of 4000, with 1cm intervals, to store the point cloud distribution within a range of ±20 meters. If the X-axis coordinate value of the truck bed point cloud data is within a certain array coordinate range, the corresponding data in that array is set to 1.

[0033] Furthermore, when determining the vehicle deflection angle, the second positioning algorithm takes point cloud data from the bottom of the truck bed 4 meters behind the vehicle, specifically the point cloud data in the Y-coordinate range (Ymax-4000, Ymax) and the Z-axis range (1000, 1600). The Z-coordinate value of the data is set to 0. The point cloud data is then subjected to a rotation base coordinate transformation at 1° intervals from -30° to 30°. The angle with the minimum value of Xmax-Xmin after rotation is taken as the candidate reference angle. The candidate reference angle ±1° is the candidate angle range. Within the candidate angle range, a rotation base coordinate transformation is performed at 0.1° intervals. The angle with the minimum value of Xmax-Xmin after rotation is taken as the vehicle deflection angle.

[0034] Furthermore, when the second positioning algorithm removes the point cloud data of the two side panels, it removes the data within the range of (Xmax-500, Xmax) and (Xmin, Xmin+500), and retains the point cloud data of the middle of the truck bed.

[0035] Furthermore, when the second positioning algorithm determines the height of the bottom of the truck bed from the ground, it takes the point cloud data within the Y-axis range (Ymax-1000, Ymax), that is, the point cloud data of 1 meter at the rear of the truck bed, and takes the data point Zmin with the smallest Z value as the height of the bottom of the truck bed from the ground.

[0036] Furthermore, when the second positioning algorithm detects the position of the vehicle's point cloud truck bed, it searches forward along the negative Y-axis from Ymax and finds data points (y1, z1) with Z coordinates greater than Zmin+300, and there are no data points less than Zmin+300 in the range (y1-400, y1-150). These data points are the position of the front baffle, and the point cloud data between the front baffle position and the distance from Ymax is the truck bed point cloud.

[0037] Furthermore, when the second positioning algorithm determines the position of the truck bed bracing, it takes data from the Y-axis range (front baffle +300, rear baffle -300) and the Z-axis range where the height of the truck bed bottom is above 300, and establishes a point cloud distribution flag array. The array size is 4000, and the point cloud distribution within a range of ±20 meters is saved at 1cm intervals. If the Y-axis coordinate value of the truck bed point cloud data is within the coordinate range of a certain array, the corresponding data in that array is set to 1.

[0038] Compared with the prior art, the beneficial technical effects of the present invention are as follows:

[0039] This invention solves the problem of locating the tie rods in the perception module of automated loading and unloading vehicles in intelligent loading and unloading robot scenarios. By acquiring the point cloud data of the truck to be tested, the first positioning algorithm is run on the two-dimensional point cloud data, and the second positioning algorithm is run on the three-dimensional point cloud data to obtain the tie rod position information. It is not limited to the two-dimensional or three-dimensional point cloud data of the vehicle, saving sensor costs, and the algorithm has a fast computing speed. It does not require calling cumbersome and complex algorithms, so it can be used in embedded and other low-resource scenarios. In addition, the trailing effect of the point cloud is processed to prevent the trailing effect from affecting the tie rod positioning. Attached Figure Description

[0040] The present invention will be further described below with reference to the accompanying drawings.

[0041] Figure 1 This is an overall flowchart of the truck body bracing positioning method based on a laser scanner according to the present invention;

[0042] Figure 2 A schematic diagram of a two-dimensional scan performed by the first laser scanner;

[0043] Figure 3 Here is the flowchart for the first localization algorithm;

[0044] Figure 4 A two-dimensional point cloud coordinate system diagram;

[0045] Figure 5 A two-dimensional point cloud map of the vehicle;

[0046] Figure 6 A two-dimensional point cloud positioning and annotation map for the vehicle;

[0047] Figure 7 A schematic diagram of a 3D scan performed by a second laser scanner;

[0048] Figure 8 Here is the flowchart for the second algorithm;

[0049] Figure 9 A 3D point cloud coordinate system diagram;

[0050] Figure 10 A 3D point cloud map of the vehicle;

[0051] Figure 11 To detect the location map of the vehicle's truck bed in the point cloud; Detailed Implementation

[0052] like Figure 1-11 As shown, a method for locating the tie rods of a truck body based on a laser scanner includes the following steps:

[0053] (1) In a rectangular testing area, a support is set up, and a first laser scanner and a pan-tilt unit are installed on the support, and a second laser scanner is installed on the pan-tilt unit;

[0054] (2) After the vehicle to be tested enters the detection area, it stops so that the first laser scanner and the second laser scanner are located on the top of the vehicle to be tested.

[0055] (3) Scan the vehicle using a first laser scanner or a second laser scanner;

[0056] (4) When the vehicle is parked at a straight angle, the first laser scanner scans the vehicle's length and height contours in a two-dimensional scan. The origin of the coordinate system is the ground directly below the first laser scanner, the X-axis points towards the rear of the vehicle, and the Y-axis points directly upward. Then, the first positioning algorithm is used, the steps of which include:

[0057] 1) Determine the height of the bottom of the truck bed from the ground. Take the point cloud data of the X-axis range (Xmax-1000, Xmax) within the detection range, that is, the point cloud data of 1 meter at the rear of the truck bed. Take the data point with the smallest Y value, Ymin, as the height of the bottom of the truck bed from the ground.

[0058] 2) Detect the position of the vehicle's point cloud truck bed. Starting from Xmax, search along the negative direction of the X-axis and find the data point (x1, y1) with a Y coordinate greater than Ymin+300. There are no data points less than Ymin+300 in the range (x1-400, x1-150). This is the position of the front baffle. The point cloud data between the front baffle position and the distance from Xmax is the truck bed point cloud.

[0059] 3) Determine the position of the truck bed tie rod, obtain point cloud data, take the X-axis range (front baffle +300, rear baffle -300) and the Y-axis range of the truck bed bottom height above 300, establish a point cloud distribution flag array, the array size is 4000, with 1cm intervals, save the point cloud distribution within ±20 meters. If the X-axis coordinate value of the truck bed point cloud data is within the coordinate range of a certain array, the corresponding data in that array is set to 1;

[0060] 4) The data of the point cloud step-by-step flag array is segmented. The adjacent distance of the X coordinate is less than 20cm and it is recorded as a segment. The start position and end position of the segment are recorded. The number of segments equals the number of tendons. The position of the segment is the position of the tendon. In the case of no tail, the difference between the start and end of the segment is the width of the tendon.

[0061] 5) The point cloud caused by the trailing only affects the ranging at the far end of the laser beam, while the ranging at the near end of the laser beam is accurate. Divide the point cloud of the truck bed into positive and negative half axes along the X-axis. The positive half axis data is based on the side with the smaller X-axis data in the segment, and the negative half axis data is based on the side with the larger X-axis data in the segment. The distance between the reference value and the maximum value of the Y-axis in the segment is multiplied by 2 to obtain the width of the beam. The width is limited to 5cm. If the beam width is greater than 5cm, then 5cm is taken as the beam width. Finally, the position and width range of the beam on the X-axis are obtained by calculation. The identified beam position is marked in the point cloud image to distinguish it from the front and rear baffles, providing obstacle avoidance positioning for the unloading port or robot gripper.

[0062] When the vehicle is parked at an angle, and the default maximum deflection angle is 30° (this default maximum deflection angle can be changed according to the actual situation), the second laser scan is performed by the gimbal, performing a three-dimensional scan. The origin of the coordinate system is the ground directly below the laser radar, the Y-axis points towards the rear of the vehicle, the X-axis points to the left side of the vehicle, and the Z-axis points directly upward. Then, the second positioning algorithm is used, the steps of which include:

[0063] 1) Determine the vehicle deflection angle. Take the point cloud data of the bottom of the truck bed 4 meters from the rear of the vehicle, that is, the point cloud data of the position in the Y coordinate range (Ymax-4000, Ymax) and the Z axis range (1000, 1600). Set the Z coordinate value of the data to 0. Perform a rotation base coordinate transformation on the point cloud data from -30° to 30° at intervals of 1°. Calculate the minimum value of Xmax-Xmin after rotation. The angle of the minimum value is the candidate reference angle. The candidate reference angle ±1° is the candidate angle range. Within the candidate angle range, perform a rotation base coordinate transformation at intervals of 0.1°. Calculate the minimum value of Xmax-Xmin after rotation. The angle of the minimum value is the vehicle deflection angle.

[0064] 2) Vehicle point cloud data correction: The vehicle point cloud data is transformed according to the deflection angle.

[0065] 3) Remove the point cloud data from both sides of the sideboards, remove the data in the range of (Xmax-500, Xmax) and (Xmin, Xmin+500), and retain the point cloud data in the middle of the truck bed;

[0066] 4) Project the point cloud data from the vehicle onto the YZ coordinate system, that is, assign a value of 0 to the X-axis of the point cloud data;

[0067] 5) Determine the height of the bottom of the truck bed from the ground. Take the point cloud data in the Y-axis range (Ymax-1000, Ymax), that is, the point cloud data of 1 meter at the rear of the truck bed. Take the data point with the smallest Z value, Zmin, as the height of the bottom of the truck bed from the ground.

[0068] 6) Detect the position of the vehicle's point cloud truck bed. Starting from Ymax, search forward along the negative Y-axis and find a data point (y1, z1) with a Z coordinate greater than Zmin+300. There are no data points less than Zmin+300 in the range (y1-400, y1-150). This is the position of the front baffle. The point cloud data between the front baffle position and the distance from Ymax is the truck bed point cloud.

[0069] 7) Determine the position of the truck bed tie rod, take the Y-axis range (front baffle +300, rear baffle -300) and the Z-axis range of truck bed bottom height above 300, establish a point cloud distribution flag array, the array size is 4000, with 1cm intervals, save the point cloud distribution within ±20 meters. If the Y-axis coordinate value of the truck bed point cloud data is within the coordinate range of a certain array, the corresponding data in that array is set to 1;

[0070] 8) The point cloud step-by-step flag array data is segmented. The adjacent distance of the Y coordinate is less than 20cm is recorded as a segment. The start position and end position of the segment are recorded. The number of segments is the number of tendons. The position of the segment is the position of the tendon. In the case of no tail, the difference between the start and end of the segment is the width of the tendon.

[0071] 9) Addressing the impact of trailing on the width of the laser beam: The laser beam has a certain divergence angle, and the spot size increases with distance. When there are two objects in front and behind, and the laser spot hits the edge of the object in front, part of the laser spot may hit the object behind. Therefore, a distance between the two is given.

[0072] 10) Process segmented data. The point cloud caused by the trailing effect only affects the ranging at the laser-far end of the tie rod, while the ranging at the laser-near end remains accurate. Divide the truck bed point cloud along the Y-axis into positive and negative halves. The positive half-axis data is based on the smaller Y-axis side within the segment, and the negative half-axis data is based on the larger Y-axis side within the segment. Multiply the distance of the Y-coordinate between the reference and the maximum Z-axis value within the segment by 2 to obtain the tie rod width. A width limit of 5cm is used; if the tie rod width is greater than 5cm, 5cm is taken as the tie rod width. Finally, the distance and width range of the tie rod on the Y-axis are calculated. The identified tie rod position is then marked on the point cloud image, distinguishing it from the front and rear baffles, to provide obstacle avoidance positioning for the unloading port or robot gripper.

[0073] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for locating the tie rods of a freight car body based on a laser scanner, characterized in that, Includes the following steps: (1) In a rectangular testing area, a support is set up, and a first laser scanner and a pan-tilt unit are installed on the support, and a second laser scanner is installed on the pan-tilt unit; (2) After the vehicle to be tested enters the detection area, it stops so that the first laser scanner and the second laser scanner are located on the top of the vehicle to be tested. (3) Scan the vehicle using a first laser scanner or a second laser scanner; (4) When the vehicle is parked at a straight angle, the first laser scanner scans the vehicle's length and height contours in a two-dimensional scan. The origin of the coordinate system is the ground directly below the first laser scanner, the X-axis points towards the rear of the vehicle, and the Y-axis points directly upward. Then, the first positioning algorithm is used, the steps of which include: 1) Determine the height of the bottom of the truck bed from the ground; 2) Detect the position of the vehicle's point cloud truck bed; 3) Determine the position of the trailer's tie rod and obtain point cloud data; 4) The data of the point cloud step-by-step flag array is segmented. The adjacent distance of the X coordinate is less than 20cm and it is recorded as a segment. The start position and end position of the segment are recorded. The number of segments equals the number of tendons. The position of the segment is the position of the tendon. In the case of no tail, the difference between the start and end of the segment is the width of the tendon. 5) The point cloud caused by the trailing only affects the ranging at the far end of the laser beam, while the ranging at the near end of the laser beam is accurate. Divide the point cloud of the truck bed into positive and negative half axes along the X-axis. The positive half axis data is based on the side with the smaller X-axis data in the segment, and the negative half axis data is based on the side with the larger X-axis data in the segment. The distance between the reference value and the maximum value of the Y-axis in the segment is multiplied by 2 to obtain the width of the beam. The width is limited to 5cm. If the beam width is greater than 5cm, then 5cm is taken as the beam width. Finally, the position and width range of the beam on the X-axis are obtained by calculation. The identified beam position is marked in the point cloud image to distinguish it from the front and rear baffles, providing obstacle avoidance positioning for the unloading port or robot gripper. When the vehicle is parked at an angle, and the default maximum deflection angle is 30° (this default maximum deflection angle can be changed according to the actual situation), the second laser scan is performed by the gimbal, performing a three-dimensional scan. The origin of the coordinate system is the ground directly below the laser radar, the Y-axis points towards the rear of the vehicle, the X-axis points to the left side of the vehicle, and the Z-axis points directly upward. Then, the second positioning algorithm is used, the steps of which include: 1) Determine the vehicle's deflection angle; 2) Vehicle point cloud data correction: The vehicle point cloud data is transformed according to the deflection angle. 3) Remove the point cloud data from both side panels; 4) Project the point cloud data from the vehicle onto the YZ coordinate system, that is, assign a value of 0 to the X-axis of the point cloud data; 5) Determine the height of the bottom of the truck bed from the ground; 6) Detect the position of the vehicle's point cloud truck bed; 7) Determine the location of the tie rods in the truck bed; 8) The point cloud step-by-step flag array data is segmented. The adjacent distance of the Y coordinate is less than 20cm is recorded as a segment. The start position and end position of the segment are recorded. The number of segments is the number of tendons. The position of the segment is the position of the tendon. In the case of no tail, the difference between the start and end of the segment is the width of the tendon. 9) Addressing the impact of trailing on the width of the laser beam: The laser beam has a certain divergence angle, and the spot size increases with distance. When there are two objects in front and behind, and the laser spot hits the edge of the object in front, part of the laser spot may hit the object behind. Therefore, a distance between the two is given. 10) Process segmented data. The point cloud caused by the trailing only affects the ranging at the far end of the laser beam, while the ranging at the near end is accurate. Divide the point cloud of the truck bed into positive and negative half-axis along the Y-axis. The positive half-axis data is based on the smaller Y-axis side within the segment, and the negative half-axis data is based on the larger Y-axis side within the segment. Multiply the distance of the Y-coordinate between the reference and the maximum value of the Z-axis within the segment by 2 to obtain the width of the beam. Set the width limit at 5cm. If the beam width is greater than 5cm, take 5cm as the beam width. Finally, calculate the distance and width range of the beam on the Y-axis. Mark the identified beam position and the front and rear baffles in the point cloud image to provide obstacle avoidance positioning for the unloading port or robot gripper.

2. The method for positioning the tie rod of a freight car body based on a laser scanner according to claim 1, characterized in that: When the first positioning algorithm determines the height of the bottom of the truck bed from the ground, it takes the point cloud data of the X-axis range (Xmax-1000, Xmax) within the detection range, that is, the point cloud data of 1 meter at the tail of the truck bed, and takes the data point with the smallest Y value, Ymin, as the height of the bottom of the truck bed from the ground. The data unit is mm if there is no label.

3. The method for positioning the tie rod of a freight car body based on a laser scanner according to claim 2, characterized in that: When the first localization algorithm detects the position of the truck bed in the vehicle point cloud, it searches along the negative direction of the X-axis from Xmax and finds data points (x1, y1) with Y coordinates greater than Ymin+300. There are no data points less than Ymin+300 in the range (x1-400, x1-150). This is the position of the front baffle. The point cloud data between the position of the front baffle and the distance of Xmax is the truck bed point cloud.

4. The method for positioning the tie rod of a freight car body based on a laser scanner according to claim 3, characterized in that: The first positioning algorithm determines the position of the truck bed tie rod, takes data within the X-axis range (front baffle +300, rear baffle -300) and the Y-axis range (the height of the truck bed bottom is above 300), and establishes a point cloud distribution flag array with an array size of 4000. The array is divided into 1cm intervals and stores the point cloud distribution within a range of ±20 meters. If the X-axis coordinate value of the truck bed point cloud data is within the coordinate range of a certain array, the corresponding data in that array is set to 1.

5. The method for positioning the tie rod of a freight car body based on a laser scanner according to claim 4, characterized in that: When determining the vehicle deflection angle, the second positioning algorithm takes point cloud data from the bottom of the truck bed 4 meters behind the vehicle, specifically the point cloud data in the Y-coordinate range (Ymax-4000, Ymax) and the Z-axis range (1000, 1600). The Z-coordinate value of the data is set to 0. The point cloud data is then subjected to a rotation base coordinate transformation at 1° intervals from -30° to 30°. The angle with the minimum value of Xmax-Xmin after rotation is taken as the candidate reference angle. The candidate reference angle ±1° is the candidate angle range. Within the candidate angle range, a rotation base coordinate transformation is performed at 0.1° intervals. The angle with the minimum value of Xmax-Xmin after rotation is taken as the vehicle deflection angle.

6. The method for positioning the tie rod of a freight car body based on a laser scanner according to claim 5, characterized in that: When the second positioning algorithm removes the point cloud data of the two side panels, it removes the data in the range of (Xmax-500, Xmax) and (Xmin, Xmin+500), and retains the point cloud data in the middle of the truck bed.

7. The method for positioning the tie rod of a freight car body based on a laser scanner according to claim 6, characterized in that: When determining the height of the bottom of the truck bed from the ground, the second positioning algorithm takes the point cloud data within the Y-axis range (Ymax-1000, Ymax), that is, the point cloud data of 1 meter at the tail of the truck bed, and takes the data point Zmin with the smallest Z value as the height of the bottom of the truck bed from the ground.

8. The method for positioning the tie rod of a freight car body based on a laser scanner according to claim 7, characterized in that: When the second positioning algorithm detects the position of the truck bed in the vehicle point cloud, it searches forward along the negative Y-axis from Ymax and finds a data point (y1, z1) with a Z coordinate greater than Zmin+300 and no data point less than Zmin+300 in the range (y1-400, y1-150). This is the position of the front baffle, and the point cloud data between the position of the front baffle and the distance of Ymax is the truck bed point cloud.

9. The method for positioning the tie rod of a freight car body based on a laser scanner according to claim 8, characterized in that: When the second positioning algorithm determines the position of the truck bed bracing, it takes data from the Y-axis range (front baffle +300, rear baffle -300) and the Z-axis range where the bottom height of the truck bed is above 300. It establishes a point cloud distribution flag array with an array size of 4000, with 1cm intervals, and saves the point cloud distribution within a range of ±20 meters. If the Y-axis coordinate value of the truck bed point cloud data is within the coordinate range of a certain array, the corresponding data in that array is set to 1.

Images