Full-automatic vehicle body pose detection system and method

Abstract

The present application relates to a kind of full automatic loading vehicle body pose detection system and method.The system includes two one-dimensional laser ranging sensors, two-dimensional laser radar, lifting platform, conveyor belt mounted on lifting platform and host computer.Two-dimensional laser radar is installed on the surface of lifting platform;Two one-dimensional laser ranging sensors are arranged at the end of conveyor belt, and the laser emitted by the two is located in the same vertical plane and opposite direction.Two-dimensional laser radar, two one-dimensional laser ranging sensors transmit the data collected to host computer.The present application combines two-dimensional laser radar and one-dimensional laser ranging sensor, realizes the automatic measurement and positioning of truck compartment size, with the characteristics of simple principle, small environmental interference influence etc.The present application scans the vehicle after stopping, calculates the length, width, height, angle of deviation and absolute position of each point of vehicle compartment, provides basic data for loading package.

Description

Technical Field

[0001] This invention relates to the field of posture detection technology, specifically to a fully automated vehicle posture detection system and method. Background Technology

[0002] Currently, material loading in China mainly relies on manual labor, which is not only slow but also involves high labor intensity and costs for workers. To address these issues and improve loading efficiency, there is an urgent need to develop a fully automated loading system to replace manual loading.

[0003] Before fully automated loading, the condition of the truck bed needs to be assessed to ensure there are no obstructions affecting loading and stacking, the dimensions of the bed need to be determined, and the loading equipment adjusted according to the bed's tilt angle for accurate loading. Car body pose detection is one of the core components of a fully automated loading system, and its accuracy directly affects the overall loading effect. Existing methods for measuring the truck bed generally fall into two categories: one uses multiple ranging sensors to acquire data, which, while simple, cannot determine the items placed inside the bed; the other uses 3D LiDAR or depth cameras to obtain information and determine the bed's position, which, while providing detailed information about the bed's interior, is more expensive, has complex algorithms, and is time-consuming. Summary of the Invention

[0004] The purpose of this invention is to provide a fully automated vehicle body pose detection system and method for vehicle loading. This detection system and method can overcome the shortcomings of the prior art, achieve high-precision pose detection, improve the accuracy of vehicle loading control and positioning, and enhance the overall loading effect.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] In a first aspect of the present invention, a fully automated vehicle body posture detection system for vehicle assembly is disclosed.

[0007] The system includes:

[0008] A one-dimensional laser rangefinder sensor is used to measure the height of a truck bed.

[0009] Two-dimensional lidar is used to measure the length, width, and deflection angle of a truck bed relative to a set position.

[0010] Furthermore, the number of one-dimensional laser ranging sensors is two, namely a first laser ranging sensor and a second laser ranging sensor.

[0011] Furthermore, the system also includes a conveyor belt;

[0012] The first laser rangefinder and the second laser rangefinder are respectively located on two relatively moving surfaces of the conveyor belt;

[0013] Driven by the main control system, the conveyor belt moves the first laser ranging sensor and the second laser ranging sensor in a straight line along the length of the truck bed.

[0014] Furthermore, the system also includes a lifting platform, on which both the conveyor belt and the two-dimensional lidar are mounted. The two-dimensional lidar is mounted on the surface of the lifting platform and is located at the center of the front end of the lifting platform; a one-dimensional laser rangefinder is located at the end of the conveyor belt, and the lasers emitted by the two one-dimensional laser rangefinders are located in the same vertical plane and in opposite directions.

[0015] Furthermore, the system also includes a host computer, and both the one-dimensional laser rangefinder and the two-dimensional lidar are connected to the host computer via serial communication. The two-dimensional lidar and the one-dimensional laser rangefinder are connected to the host computer and transmit the collected data to the host computer.

[0016] In a second aspect of the present invention, a detection method for the above-mentioned detection system is disclosed, which is used to measure and position the cargo compartment of a freight car.

[0017] The method includes:

[0018] S1. Use a two-dimensional lidar to obtain point cloud data of the carriage's XOY plane and send the data to the host computer.

[0019] S2. Preprocess the point cloud data;

[0020] S3. Based on the preprocessed point cloud data, determine whether there are any items inside the vehicle; if there are no items inside the vehicle, proceed to step S4; if there are items inside the vehicle, return to step S1.

[0021] S4. Based on the preprocessed point cloud data, determine the length, width, deflection angle, and stopping position of the carriage;

[0022] S5. Determine the height of the carriage using data collected by a one-dimensional laser rangefinder sensor.

[0023] Furthermore, the preprocessing of the point cloud data includes:

[0024] S21. Determine the boundary points of the carriage based on the point cloud data;

[0025] S22. Perform coordinate transformation on the point cloud data;

[0026] S23. The point cloud data is smoothed by using the mean filtering method.

[0027] Further, based on the preprocessed point cloud data, it is determined whether there are items in the vehicle; if there are no items in the carriage, step S4 is executed; if there are items in the carriage, return to step S1, including:

[0028] S31. Connect a line between the first and last points of the data point set in the preprocessed point cloud data;

[0029] S32. Calculate the distances from all points to the line obtained in step S31, and find the point where the maximum distance is located. The maximum distance is denoted as D max ;

[0030] S33. Compare D max with the threshold D. If D max < D, it means that all data points in this data point set belong to the same straight line, and the data processing is completed;

[0031] S34. If D max ≥ D, then determine this data point as a corner point, use this point as a segmentation point to divide the data into two data point sets, and calculate these two segments respectively according to steps S31 - S33;

[0032] S35. After all data processing is completed, obtain the data point set of all corner points. If the number of corner points in the data point set of corner points is greater than the set threshold, it is determined that there are items in the vehicle, return to execute step S1. If the number of corner point data in the data point set of corner points is not greater than the set threshold, it is determined that there are no items in the vehicle, and execute step S4.

[0033] Further, based on the preprocessed point cloud data, determining the length, width, deflection angle, and parking position of the carriage, including:

[0034] S41. Divide the acquired set of right-angle coordinate points of the carriage boundary by the corner points into multiple groups, and perform least squares linear fitting on each group respectively to obtain multiple fitting lines;

[0035] S42. Calculate the deflection angle of the carriage based on the fitted line, that is, the angle of inclination of the carriage;

[0036] S43. Calculate the length and width of the carriage based on the intersection points of the carriage boundary points and the fitted lines.

[0037] Further, determining the height of the carriage using the data collected by the one-dimensional laser distance sensor, including:

[0038] S51. If the vehicle to be measured is a closed carriage, drive the conveyor belt to move inside the carriage, start two one-dimensional laser distance sensors for measurement, obtain the distances of the two one-dimensional laser distance sensors from the vehicle floor, and determine the height of the closed carriage based on the readings of the two one-dimensional laser distance sensors and the distance between the two sensors;

[0039] S52. If the vehicle to be tested is a convertible, first raise the lifting platform to the specified height, then drive the conveyor belt to move into the interior of the vehicle, activate the first laser rangefinder to measure the distance between the first laser rangefinder and the vehicle floor, and then move the conveyor belt to the front of the vehicle and activate the first laser sensor to measure the distance between the first laser rangefinder and the roof. Based on the distance between the first laser rangefinder and the vehicle floor and the distance between the first laser rangefinder and the roof, determine the height of the convertible.

[0040] Compared with the prior art, the advantages of the present invention are:

[0041] (1) This invention combines a two-dimensional lidar and two one-dimensional laser ranging sensors to achieve automated measurement and positioning of the dimensions of the truck bed. It features simple principle and minimal impact from environmental interference. After the vehicle to be loaded is parked, the detection system in this invention scans the vehicle and calculates the length, width, height, skew angle, and absolute position of each point in the truck bed, providing basic data for loading and packing.

[0042] (2) For freight cars with obstacles inside the carriage, the present invention can make a judgment to prevent accidents from occurring during the subsequent automatic loading process.

[0043] (3) This invention covers various vehicle types, whether it is a box truck or a high-sided truck, and can measure the height of the cargo box. Attached Figure Description

[0044] Figure 1 This is a schematic diagram of the structure of the fully automated vehicle body posture detection system in this invention;

[0045] Figure 2 This is a flowchart illustrating the detection method of the fully automated vehicle body posture detection system in this invention.

[0046] Figure 3 This is the coordinate plane diagram of the linear equation used to calculate the deflection angle in this invention;

[0047] Figure 4 This is a schematic diagram showing the height measurement parameters of a closed carriage;

[0048] Figure 5 This is a schematic diagram showing the height measurement parameters of a convertible.

[0049] in:

[0050] 1. First laser rangefinder sensor, 2. Second laser rangefinder sensor, 3. Two-dimensional lidar, 4. Lifting platform, 5. Conveyor belt, 6. Host computer. Detailed Implementation

[0051] The present invention will be further described below with reference to the accompanying drawings:

[0052] like Figure 1 The fully automated vehicle body posture detection system shown includes two one-dimensional laser rangefinders (first laser rangefinder 1 and second laser rangefinder 2), a two-dimensional lidar 3, a lifting platform 4, a conveyor belt 5 mounted on the lifting platform, and a host computer 6. The two-dimensional lidar 3 is mounted on the surface of the lifting platform 4, located at the center of the front end of the platform 4. The two one-dimensional laser rangefinders 1 and 2 are located at the end of the conveyor belt 5, with the lasers emitted by the first laser rangefinder 1 and the second laser rangefinder 2 lying in the same vertical plane and in opposite directions. The two-dimensional lidar 3, the first laser rangefinder 1, and the second laser rangefinder 2 are connected to the host computer 6 and transmit the collected data to the host computer 6.

[0053] Figure 2 This is a flowchart illustrating the detection method of the fully automated vehicle body pose detection system of the present invention. The detection method includes:

[0054] S1. Use a two-dimensional lidar to obtain point cloud data of the carriage's XOY plane and send the data to the host computer.

[0055] For applications applicable to this invention, the rear side panel of the truck bed should be open, allowing the conveyor belt to enter from the rear of the truck bed for loading and stacking during the loading process. After the vehicle to be tested moves to the designated area, due to the different floor heights of different truck beds, the lifting platform needs to be raised to a preset height so that the two-dimensional lidar can acquire point cloud data of the truck bed interior. The acquired point cloud data is the data from one complete radar scan, specifically a polar coordinate data point set {(θ1, d1), (θ2, d2), ..., (θ...}. n d n ), …, (θ m d m )}, where θ represents the angle between the measurement point and the radar itself, d represents the distance between the measurement point and the radar, and θ n d represents the angle value of the nth data point acquired by the radar. n This represents the distance value of the nth data point acquired by the radar, and m represents the total number of data points acquired by the radar in one scan cycle.

[0056] S2. Preprocess the point cloud data.

[0057] The preprocessing of point cloud data includes:

[0058] S21. Determine the boundaries of the carriage, obtain the two boundary points of the carriage, extract all data between the two points, and perform the next calculation.

[0059] Because the carriage has no rear panel, the data value will suddenly increase when the laser scan reaches the carriage boundary. Based on this condition, we can determine that this is the carriage boundary. We then obtain two boundary points A(θ) of the carriage. a d a ), B(θ) b d b Extract all data between two points to obtain a new polar coordinate point set {(θ)}. a d a ), (θ a+1 d a+1 ), …, (θ n d n ), …, (θ b d b )}, and proceed to the next calculation.

[0060] S22. Perform coordinate transformation on the data.

[0061] The data returned by a 2D lidar is rotation angle and distance data. For easier subsequent calculations, it needs to be converted to Cartesian coordinates. The 2D lidar returns rotation angle θ and distance d, and the coordinates of this point in the Cartesian coordinate system are (d·cosθ, d·sinθ). This yields a new set of Cartesian coordinate points {(x...}... a y a ), (x a+1 y a+1 ), ..., (x n y n ), ..., (x b y b )}, where (x n y n ) corresponds to (d n ·cosθ n d n ·sinθ n ).

[0062] S23, Smoothing filter processing.

[0063] Due to the influence of external environment, instrument inherent errors, and measurement errors, 2D LiDAR suffers from various errors during measurement. Without denoising and smoothing the data, these noise points directly affect the accuracy of feature point extraction, resulting in significant measurement errors. Therefore, denoising and smoothing of the point cloud data is necessary. A mean filtering method is used to process the data, replacing the value of each point in the data set with the mean of adjacent data points within a certain range, thereby eliminating large fluctuations in data noise. After processing, a new rectangular coordinate point set {(x'... a y' a ), (x'a+1 , y’ a+1 ), …, (x’ n , y’ n ), …, (x’ b , y’ b )}。

[0064] S3. Based on the preprocessed point cloud data, determine whether there are items in the vehicle; if there are no items in the carriage, execute step S4; if there are items in the carriage, return to step S1.

[0065] During the observation of the carriage, some corners will be obtained for the vertices of the carriage and the places where items are placed. Such corners are called corner points. If there are no items placed in the carriage, the corner point detection result is 4; if the final corner point detection result is greater than 4, it means there are items in the vehicle and cargo placement cannot be carried out. The specific steps are as follows:

[0066] S31. Connect a line between the first and last two points of the data point set in the preprocessed point cloud data.

[0067] S32. Calculate the distances from all points to the straight line obtained in step S31, and find the point where the maximum distance is located. The maximum distance is denoted as D max .

[0068] S33. Compare D max with the threshold D. If D max < D, it means that all data points in this data point set belong to the same straight line and the data processing is completed.

[0069] S34. If D [[ID=ON]]t max >= D, then determine this data point as a corner point, use this point as a segmentation point, divide the data into two data point sets, and calculate each of these two segments according to steps S31 - S33.

[0070] S35. After all data processing is completed, obtain the data point set of all corner points. If the number of corner points in the data point set of corner points is greater than the set threshold, it is determined that there are items in the vehicle, return to execute step S1; if the number of corner points in the data point set of corner points is not greater than the set threshold, it is determined that there are no items in the vehicle, and execute step S4. In this embodiment, if the corner point detection result is 4, then execute step S4; if the final corner point detection result is greater than 4, it means there are items in the vehicle and cargo placement cannot be carried out, then return to step S1 and wait for the staff to move the items away and then re - detect.

[0071] S4. Based on the preprocessed point cloud data, determine the length, width, deflection angle, and docking position of the carriage.

[0072] The determination of the length, width, deflection angle, and docking position of the carriage based on the preprocessed point cloud data includes:

[0073] S41. Divide the obtained rectangular coordinate points of the carriage boundary into three groups, using the corner points as boundaries. Perform least squares linear fitting on each group to obtain three fitted lines {L1, L2, L3}. Among them, L1 and L3 are parallel and represent the line containing the length of the carriage; L2 is perpendicular to L1 and L3 and represents the line containing the width of the carriage.

[0074] S42. Calculate the car body deflection angle, i.e., the angle at which the car body tilts, based on the fitted straight line. The deflection angle refers to the rotation angle between the length side of the car body and the Y-axis, defined as α. There are two possibilities for the angle α between the length side of the car body and the Y-axis: one is that the car body deflects in the positive direction of the X-axis, and the other is that it deflects in the negative direction of the X-axis. The two deflection cases are as follows: Figure 2 As shown.

[0075] Given the slope *a* of the straight line containing the long side of the carriage, and the angle θ between the long side of the carriage and the positive X-axis = arctan(a), then the rotation angle α = 90° - θ. From... Figure 3 As can be seen, when the calculated angle α is greater than 0°, the carriage deviates in the positive direction of the X-axis by an angle of α; when the calculated angle α is less than 0°, the carriage deviates in the negative direction of the X-axis by an angle of -α.

[0076] S43. The length and width of the carriage are obtained from the intersection of the boundary points and the fitted straight line.

[0077] Given the boundary point A(x) of the carriage a y a ), B(x) b y b Based on the three lines calculated in step S41, the intersection point C(x) of L1 and L2 is obtained. c y c The intersection point D(x) of L3 and L2 d y d The length and width of the carriage can be determined based on the four points A, B, C, and D.

[0078] The calculation process for the width of the carriage is as follows:

[0079] The distance W between A and B AB for:

[0080] The distance W between C and D CD for:

[0081] Therefore, the width W of the carriage is:

[0082] Similarly, the length L of the carriage can be obtained as:

[0083] S5. The one-dimensional laser rangefinder sends the measurement data to the host computer, which then determines the height of the carriage based on the data collected by the one-dimensional laser rangefinder.

[0084] S51. After the vehicle under test moves to the designated area, if the vehicle is a closed compartment, the drive conveyor belt moves forward 2 meters until it reaches the interior of the compartment. Then, two one-dimensional laser rangefinders are activated to perform measurements. Figure 4 As shown, the distance H1 between the first laser rangefinder and the vehicle floor is obtained, and the distance H2 between the second laser rangefinder and the vehicle roof is obtained. Based on the readings of the two one-dimensional laser rangefinders and the distance d between the two sensors, the height of the enclosed vehicle compartment can be calculated, that is, the height H of the vehicle compartment is H = H1 + H2 + d.

[0085] S52. After the vehicle under test is moved to the designated area, if the vehicle is an open-top carriage, first raise the lifting platform to the designated height, then drive the conveyor belt forward 2 meters to reach the interior of the carriage. Figure 5 As shown in (a), the first laser rangefinder is activated to measure and obtain the distance H1 between the first laser rangefinder and the vehicle floor. Continue moving to the front of the vehicle, as shown... Figure 5 As shown in (b), the first laser sensor is activated to measure and obtain the distance H2 between the first laser rangefinder and the roof of the vehicle. Based on the two readings of the first laser rangefinder, the height of the convertible can be calculated, that is, the height H of the canopy is H = H1 - H2.

[0086] The above-described embodiments 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 fully automated vehicle body posture detection system for vehicle assembly, characterized in that, The system includes: A one-dimensional laser ranging sensor for measuring the height of the freight car compartment; A two-dimensional lidar for measuring the length, width of the freight car compartment and the deflection angle relative to the set position; The number of the one-dimensional laser ranging sensors is two, namely the first laser ranging sensor and the second laser ranging sensor; The system further includes a conveyor belt; The first laser ranging sensor and the second laser ranging sensor are respectively located on two opposite moving surfaces of the conveyor belt; Driven by the main control system, the conveyor belt drives the first laser ranging sensor and the second laser ranging sensor to linearly move along the length direction of the freight car compartment; The system further includes a lifting platform, and both the conveyor belt and the two-dimensional lidar are installed on the lifting platform; The detection method of the system includes: S1. Obtain the point cloud data of the compartment in the XOY plane by using the two-dimensional lidar; S2. Preprocess the point cloud data; S3. According to the preprocessed point cloud data, judge whether there are items in the vehicle; if there are no items in the compartment, execute step S4; if there are items in the compartment, return to step S1, including: S31. Connect a line between the first and last two points of the data point set in the preprocessed point cloud data; S32. Calculate the distance from all points to the line obtained in step S31, and find the point where the maximum distance is located. The maximum distance is denoted as Dmax; S33. Compare Dmax with the threshold D. If Dmax < D, it means that all data points in this data point set belong to the same straight line, and the data processing is completed; S34. If Dmax >= D, then determine this data point as a corner point, use this data point as a segmentation point, divide the data into two data point sets, and calculate these two segments respectively according to steps S31~S33; S35. After all data processing is completed, obtain the data point set of all corner points. If the number of corner points in the data point set of corner points is greater than the set threshold, it is determined that there are items in the vehicle, return to execute step S1. If the number of corner point data in the data point set of corner points is not greater than the set threshold, it is determined that there are no items in the vehicle, and execute step S4; S4. Determine the length, width, deflection angle and docking position of the compartment according to the preprocessed point cloud data; S5. Determine the height of the compartment by using the data collected by the one-dimensional laser ranging sensor.

2. The system according to claim 1, wherein The system further includes a host computer, and both the one-dimensional laser ranging sensor and the two-dimensional lidar are connected to the host computer through serial communication.

3. The system according to claim 1, characterized in that, The preprocessing of the point cloud data includes: S21. Determine the boundary points of the compartment according to the point cloud data; S22. Perform coordinate transformation on the point cloud data; S23. Perform smoothing filtering on the point cloud data by using the mean filtering method.

4. The system according to claim 1, wherein The determining the length, width, deflection angle and docking position of the compartment according to the preprocessed point cloud data includes: S41. Divide the obtained set of rectangular coordinate points of the carriage boundary into multiple groups with the corner points as the boundaries, and perform least squares linear fitting on each group to obtain multiple fitted lines. S42. Calculate the deflection angle of the carriage, i.e. the angle at which the carriage tilts, based on the fitted straight line. S43. The length and width of the carriage are obtained from the intersection of the carriage boundary points and the fitted straight line.

5. The system according to claim 4, characterized in that, The process of determining the height of the carriage using data collected by a one-dimensional laser rangefinder includes: S51. If the vehicle to be tested is a closed compartment, drive the conveyor belt to move into the compartment, start two one-dimensional laser rangefinders to measure, obtain the distances of the two one-dimensional laser rangefinders from the floor and roof of the vehicle, and determine the height of the closed compartment based on the readings of the two one-dimensional laser rangefinders and the distance between the two sensors. S52. If the vehicle to be tested is a convertible, first raise the lifting platform to the specified height, then drive the conveyor belt to move into the interior of the vehicle, activate the first laser rangefinder to measure the distance between the first laser rangefinder and the vehicle floor, and then move the conveyor belt to the front of the vehicle and activate the first laser sensor to measure the distance between the first laser rangefinder and the roof. Based on the distance between the first laser rangefinder and the vehicle floor and the distance between the first laser rangefinder and the roof, determine the height of the convertible.

Images