A single-line laser radar tire detection method based on feature evaluation function

Abstract

The present invention belongs to the technical field of tire detection, and specifically relates to a single-line laser radar tire detection method based on a feature evaluation function. Use a single-line laser radar device to scan the surface of a vehicle tire to obtain original three-dimensional point cloud data, wherein the point cloud data includes the three-dimensional coordinates of each point and the reflection intensity information at the point; pre-process the obtained original three-dimensional point cloud data to obtain pre-processed three-dimensional point cloud data; based on the pre-processed three-dimensional point cloud data, obtain tire size characteristics; based on the pre-processed three-dimensional point cloud data, perform reflection intensity threshold filtering to obtain tire strength characteristics; based on the pre-processed three-dimensional point cloud data, use a motion compensation algorithm to obtain motion characteristics; based on the tire size characteristics, tire strength characteristics and motion characteristics, design a feature evaluation function to obtain final tire information. By using the designed evaluation function, wheels can be effectively identified and detection accuracy can be improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of tire detection, and in particular relates to a single-line laser radar tire detection method based on a characteristic evaluation function. Background Art

[0002] In the field of automobile manufacturing and logistics, commercial vehicle transfer robots have become key equipment for improving production efficiency, reducing labor costs and ensuring operational safety. These robots realize automatic handling and transfer of commercial vehicles between different production stages or logistics nodes through precise control. In this process, precise positioning of the center of the commercial vehicle wheel is one of the core technologies to ensure accurate and safe operation of the robot.

[0003] Traditional LiDAR-based tire detection methods rely on the collection and analysis of point cloud data; however, this method may face the problem of reduced recognition accuracy in complex environments (such as multi-vehicle interference, occlusion, tire stains, etc.), and it is difficult to fully evaluate key features such as tire size, strength and motion status.

[0004] Although visual sensors can provide rich image information, their performance is highly dependent on environmental conditions. Light changes, bad weather, shadows, and camera viewing angle limitations can affect the accuracy of wheel detection. In addition, the image processing algorithm is complex and requires large computing resources, which may affect real-time performance.

[0005] Whether it is a lidar or a visual sensor, the data it acquires needs to go through complex preprocessing, feature extraction, target recognition and positioning steps. These processes not only increase the complexity of the algorithm, but may also introduce additional error sources. At the same time, in order to meet real-time requirements, the algorithm must have efficient computing capabilities while ensuring accuracy.

[0006] In the dynamic environment of automobile manufacturing and logistics, information such as vehicle position, speed and direction needs to be updated in real time to support the precise control and decision-making of the transfer robot. Therefore, the wheel detection algorithm must have high real-time performance and be able to complete data processing and result output in a very short time. Summary of the invention

[0007] In order to overcome the problems in the prior art, the present invention proposes a single-line laser radar tire detection method based on a feature evaluation function.

[0008] The technical solution of the present invention to solve the above technical problems is as follows:

[0009] The present invention provides a single-line laser radar tire detection method based on a feature evaluation function, comprising the following steps:

[0010] Step 1: Use a single-line laser radar device to scan the surface of the vehicle tire to obtain original three-dimensional point cloud data, where the point cloud data includes the three-dimensional coordinates of each point and the reflection intensity information at the point;

[0011] Step 2: preprocessing the obtained original three-dimensional point cloud data to obtain preprocessed three-dimensional point cloud data;

[0012] Step 3: Obtain tire size characteristics based on the preprocessed 3D point cloud data;

[0013] Step 4: Based on the preprocessed 3D point cloud data, perform reflection intensity threshold filtering to obtain tire strength characteristics;

[0014] Step 5: Based on the preprocessed 3D point cloud data, a motion compensation algorithm is used to obtain motion features;

[0015] Step 6: Based on tire size characteristics, tire strength characteristics and motion characteristics, design a feature evaluation function to obtain final tire information.

[0016] Furthermore, in step 2, the original three-dimensional point cloud data is preprocessed, including removing outliers based on Gaussian filtering to obtain preprocessed three-dimensional point cloud data.

[0017] Furthermore, in step 3, obtaining tire size characteristics based on the preprocessed three-dimensional point cloud data includes:

[0018] Obtain a chord of the tire through a single-line laser radar L , that is, the part where the laser beam intersects the tire, combined with the laser radar installation height H， Calculate tire size information:

[0019] ;

[0020] In the above formula, Indicates the size information of the tire, that is, the diameter information of the tire.

[0021] Furthermore, in step 4, based on the pre-processed three-dimensional point cloud data, reflection intensity threshold filtering is performed to obtain tire strength characteristics, which specifically includes:

[0022] Traverse each point in the preprocessed 3D point cloud data and check its reflection intensity value;

[0023] If the reflection intensity value of a certain point is lower than the preset reflection intensity threshold , then this point is regarded as a noise point and filtered out;

[0024] If the reflection intensity value of a certain point is greater than or equal to the preset reflection intensity threshold , then this point is retained, and the average intensity of the points whose reflection intensity is greater than or equal to the threshold is calculated, recorded as I。

[0025] Furthermore, in step 5, based on the preprocessed three-dimensional point cloud data, a motion compensation algorithm is used to obtain motion features, which specifically includes:

[0026] The ICP-based motion compensation algorithm solves the rigid body transformation relationship between two point clouds by minimizing the distance between them, including:

[0027] Step 51: Set the target function; top of the form

[0028] Suppose there are two point cloud datasets, source point cloud P and target point cloud Q, containing n and m points respectively, where , ;

[0029] ;

[0030] In the above formula, T represents the rigid body transformation matrix, E represents the objective function, represents a point in the source point cloud P, is the corresponding point in the target point cloud Q;

[0031] Step 52: Corresponding point search: The ICP algorithm needs to find corresponding point pairs in two point clouds;

[0032] For a point in the source point cloud P, find the nearest point in the target point cloud Q as the corresponding point, which is achieved by calculating the Euclidean distance between all point pairs and taking the minimum value;

[0033] Step 53: Use the least square method to solve the rotation matrix and the translation vector to minimize the objective function;

[0034] Step 54: Repeat the above steps, and each iteration uses the transformation obtained in the previous iteration to update the position of the source point cloud P, and re-search the corresponding points; the iteration continues until the preset stop condition is met, and the tire center position at time k is obtained. .

[0035] Furthermore, in step 6, based on the tire size characteristics, tire strength characteristics and motion characteristics, a characteristic evaluation function is designed to convert the cost value The smallest target is used as the final target with the highest confidence:

[0036] ;

[0037] in, is the standard strength value of the target commercial vehicle tire, is the standard diameter value of the target commercial vehicle tire, yes k -1 The position of the tire center at the moment, is the position of the tire center at time k; I represents the point average intensity, Indicates tire size information.

[0038] Compared with the prior art, the present invention has the following technical effects:

[0039] The present invention extracts specific tire parameters from lidar data, comprehensively considers characteristics such as size, strength and movement speed, and designs a characteristic evaluation function, which can improve the accuracy and efficiency of commercial vehicle tire positioning, thereby increasing the safety and performance of the transfer robot's transfer operation on the target vehicle. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0041] Figure 1 It is a schematic diagram of the process of the present invention;

[0042] Figure 2 A schematic diagram of tire size characteristics is obtained for the present invention. DETAILED DESCRIPTION

[0043] In order to further explain the technical means and effects taken by the present invention to achieve the predetermined invention purpose, the specific implementation methods, structures, features and effects of the technical solutions proposed by the present invention are described in detail below in conjunction with the accompanying drawings and preferred embodiments. The specific features, structures or characteristics in one or more embodiments may be combined in any suitable form. Unless otherwise defined, all technical and scientific terms used in the present invention have the same meaning as those commonly understood by technicians in the technical field of the present invention.

[0044] Reference Figure 1 In one embodiment of the present invention, a single-line laser radar tire detection method based on a feature evaluation function is provided, comprising the following steps:

[0045] Step 1: Use a single-line laser radar device to scan the surface of the vehicle tire to obtain original three-dimensional point cloud data, where the point cloud data includes the three-dimensional coordinates of each point and the reflection intensity information at the point;

[0046] Step 2: preprocessing the obtained original three-dimensional point cloud data to obtain preprocessed three-dimensional point cloud data;

[0047] Step 3: Obtain tire size characteristics based on the preprocessed 3D point cloud data;

[0048] Step 4: Based on the preprocessed 3D point cloud data, perform reflection intensity threshold filtering to obtain tire strength characteristics:

[0049] Step 5: Based on the preprocessed 3D point cloud data, a motion compensation algorithm is used to obtain motion features;

[0050] Step 6: Based on tire size characteristics, tire strength characteristics and motion characteristics, design a feature evaluation function to obtain final tire information.

[0051] The following is a detailed explanation of each of the above steps:

[0052] Step 1: Use a single-line laser radar device to scan the surface of the vehicle tire to obtain original 3D point cloud data;

[0053] A single-line laser radar device is used to scan the surface of the vehicle tire to obtain original three-dimensional point cloud data. The point cloud data includes the three-dimensional coordinates of each point and the reflection intensity information at the point, which is usually related to the energy of the laser beam reflected back at the point.

[0054] As an example, this step may include:

[0055] Step 11: Obtain a series of scans through a single-line laser radar, each scan including distance and angle information of multiple laser beams, and resolve the distance and angle information of each laser beam;

[0056] Mount the single-line LiDAR on a robot or fixed bracket to ensure that the laser beam can scan the surface of the vehicle tire.

[0057] Start the single-line laser radar and perform continuous scanning. The single-line laser radar obtains the SCAN data type, that is, the single-line laser radar usually outputs a series of scans (SCAN), which are stored in a data packet (pcap) in binary form; each scan includes the distance and angle information of multiple laser beams, which are emitted at different vertical angles and measure the distance to the nearest obstacle. Specifically, subscribe to LaserScan messages: Based on the ROS node, subscribe to the SCAN information from the laser radar.

[0058] Parse the data packet, parse the distance value and angle information of each laser beam, and extract the distance values ​​measured at different angles; by acquiring the point cloud data, the point cloud distribution of the entire yard can be obtained, which is convenient for subsequent processing.

[0059] Step 12: Convert the acquired distance and angle information of the laser beam into original three-dimensional point cloud data.

[0060] Convert the distance data into three-dimensional point cloud data, where each point contains three-dimensional coordinates;

[0061] Assume that the lidar is located at the origin (0,0,0);

[0062] ;

[0063] in, is the distance data, indicating the distance from the origin; is the elevation angle, the angle from the positive Z axis (upward); is the azimuth, the angle between the line extending counterclockwise from the positive x-axis to the point and the positive x-axis.

[0064] Step 2: preprocessing the obtained original three-dimensional point cloud data, based on the preprocessed three-dimensional point cloud data;

[0065] Considering the laser radar coordinate system, the reflection intensity information of the tire surface is extracted from the laser radar data. The Gaussian filtering algorithm is used to remove noise and outliers. Through intensity filtering, the data quality is improved to ensure the accuracy of subsequent processing.

[0066] Preprocess the original 3D point cloud data and remove outliers based on Gaussian filtering; Gaussian filtering function for:

[0067] ;

[0068] in, x is the value of the point cloud data, is the mean, is the standard deviation.

[0069] Step 3: Obtain tire size characteristics based on the preprocessed 3D point cloud data;

[0070] Reference Figure 2 , a single-line laser radar obtains a chord of the tire L , that is, the part where the laser beam intersects the tire, combined with the laser radar installation height H .

[0071] The installation position of the single-line laser radar allows the laser beam to irradiate the tire. When the tire is perpendicular to the laser beam, the following can be obtained according to calculation:

[0072] ;

[0073] Based on this, the tire size information, that is, the tire diameter information, can be obtained, which is recorded as .

[0074] Step 4: Based on the preprocessed 3D point cloud data, perform reflection intensity threshold filtering to obtain tire strength characteristics;

[0075] Preset reflection intensity threshold , reflection intensity threshold The selection is based on the reflective characteristics of the target object, i.e. the tire, the ambient lighting conditions, and the expected noise level setting. The purpose is to retain the high-intensity reflection points originating from the tire and remove the low-intensity noise points caused by the background, ground or other non-target sources.

[0076] Traverse each point in the point cloud dataset and check its reflection intensity value. If the reflection intensity value of a point is lower than the preset reflection intensity threshold , then this point is regarded as a noise point and filtered out; if the reflection intensity value of a point is greater than or equal to the preset reflection intensity threshold , then this point is retained, and the average intensity of the points whose reflection intensity is greater than or equal to the threshold is recorded as I , these points correspond to the signals reflected from the tire surface. By setting the intensity threshold , the reflection intensity information of the tire can be extracted from the original laser point cloud data, and the noise points with lower intensity can be filtered out.

[0077] Step 5: Based on the preprocessed 3D point cloud data, a motion compensation algorithm is used to obtain motion features;

[0078] When the transfer robot is driving, a motion compensation algorithm is used to correct errors in the lidar scanning data caused by the robot's acceleration, deceleration, turning, etc., to ensure accurate identification and positioning of the commodity vehicle.

[0079] The mathematical expression of the ICP (Iterative Closest Point) algorithm mainly revolves around minimizing a certain distance metric between two point clouds, usually the sum of squared Euclidean distances. This optimization problem can be formalized as finding an optimal rigid body transformation matrix T, which can include a rotation matrix and a translation vector, so that the distance between the corresponding point pairs between the transformed point cloud and the target point cloud is minimized.

[0080] The ICP-based motion compensation algorithm solves the rigid body transformation relationship between two point clouds by minimizing the distance between them, including:

[0081] Step 51: Set the target function; top of the form

[0082] Suppose there are two point cloud datasets, source point cloud P and target point cloud Q, containing n and m points respectively, where , ;

[0083] The goal of the ICP algorithm is to find a rigid body transformation matrix T so that the points in the source point cloud P Corresponding points in the target point cloud Q closest, It can be achieved by minimizing the following objective function E To achieve:

[0084] ;

[0085] Step 52: Corresponding point search: The ICP algorithm needs to find corresponding point pairs in two point clouds;

[0086] The nearest point can be searched by Euclidean distance. For each point in a point cloud, the nearest point can be found in another point cloud as the corresponding point.

[0087] For each point in the source point cloud P, find the nearest point in the target point cloud Q as the corresponding point, which is usually achieved by calculating the Euclidean distance between all point pairs and taking the minimum value.

[0088] Step 53: Use the least square method to solve the rotation matrix and the translation vector to minimize the objective function;

[0089] Based on the corresponding point pairs found, the rigid body transformation relationship between the two point clouds, i.e. translation and rotation, is estimated. The ICP algorithm usually uses the least squares method to solve this problem and determines the optimal rigid body transformation by minimizing the distance between corresponding points.

[0090] Step 54: Repeat the above steps, and each iteration uses the transformation obtained in the previous iteration to update the position of the source point cloud P, and re-searches the corresponding points; the iteration continues until a preset stop condition is met, such as the change in the objective function is less than a certain threshold, or the maximum number of iterations is reached.

[0091] Based on the motion-compensated point cloud, the predicted position of the tire center is obtained, and the tire center position is calculated using the existing algorithm. .

[0092] Step 6: Based on tire size characteristics, tire strength characteristics and motion characteristics, design a feature evaluation function to obtain final tire information.

[0093] Define a feature evaluation function that comprehensively considers the size, strength, and speed of the tire. Considering the interference of other objects in a complex environment, the positioning algorithm may misidentify "false positive targets".

[0094] A multi-dimensional feature calculation method based on tire size, reflection intensity, and motion characteristics is designed to identify and select a given tire.

[0095] Based on the reflection intensity characteristics, three-dimensional size and motion attributes of the target as feature operators, the cost function of confidence calculation is designed As shown below, it comprehensively considers the average intensity I of the point cloud and the size information of the tire and location Three factors of continuity will represent the value of The smallest target is used as the final target with the highest confidence:

[0096] ;

[0097] in, is the standard strength value of the target commercial vehicle tire, is the standard diameter value of the target commercial vehicle tire, yes k -1 The position of the tire center at the moment, is the position of the tire center at time k.

[0098] Based on the target information of the minimum cost value, which serves as the main information for the commercial vehicle transfer robot to identify and locate the wheels, the center position and size information of the commercial vehicle tires can be accurately located, and the positioning process has strong robustness and stability.

[0099] The present invention relates to single-line laser radar to obtain tire size, intensity filtering, motion compensation and feature evaluation function design. By comprehensively evaluating tire size, strength and motion speed, the method can improve vehicle safety and performance.

[0100] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.

Claims

1. A single-line laser radar tire detection method based on a feature evaluation function, characterized in that: The following steps are involved: Step 1: Use a single-line laser radar device to scan the surface of the vehicle tire to obtain original three-dimensional point cloud data, where the point cloud data includes the three-dimensional coordinates of each point and the reflection intensity information at the point; Step 2: preprocessing the obtained original three-dimensional point cloud data to obtain preprocessed three-dimensional point cloud data; Step 3: Obtain tire size characteristics based on the preprocessed 3D point cloud data; Step 4: Based on the preprocessed 3D point cloud data, perform reflection intensity threshold filtering to obtain tire strength characteristics; Step 5: Based on the preprocessed 3D point cloud data, a motion compensation algorithm is used to obtain motion features; Step 6: Based on the tire size characteristics, tire strength characteristics and motion characteristics, a characteristic evaluation function is designed to obtain the final tire information; In step 3, obtaining tire size characteristics based on the preprocessed three-dimensional point cloud data includes: Obtain a chord of the tire through a single-line laser radar L , that is, the part where the laser beam intersects the tire, combined with the laser radar installation height H， Calculate tire size information: ； In the above formula, Indicates the size information of the tire, that is, the diameter information of the tire; In step 4, based on the pre-processed three-dimensional point cloud data, reflection intensity threshold filtering is performed to obtain tire strength characteristics, which specifically includes: Preset reflection intensity threshold , reflection intensity threshold The selection is based on the reflective characteristics of the target object, i.e., the tire, the ambient lighting conditions, and the expected noise level setting. The purpose is to retain the high-intensity reflection points originating from the tire and remove the low-intensity noise points generated by the background, ground, or other non-target sources; Traverse each point in the preprocessed 3D point cloud data and check its reflection intensity value; If the reflection intensity value of a certain point is lower than the preset reflection intensity threshold , then this point is regarded as a noise point and is filtered out; If the reflection intensity value of a certain point is greater than or equal to the preset reflection intensity threshold , then this point is retained, and the average intensity of the points whose reflection intensity is greater than or equal to the threshold is calculated, recorded as I ; In step 5, based on the preprocessed three-dimensional point cloud data, a motion compensation algorithm is used to obtain motion features, which specifically includes: The ICP-based motion compensation algorithm solves the rigid body transformation relationship between two point clouds by minimizing the distance between them, including: Step 51: Setting the objective function; Suppose there are two point cloud datasets, source point cloud P and target point cloud Q, containing n and m points respectively, where , ; ； In the above formula, T represents the rigid body transformation matrix, E represents the objective function, represents a point in the source point cloud P, is the corresponding point in the target point cloud Q; Step 52: Corresponding point search: The ICP algorithm needs to find corresponding point pairs in two point clouds; For a point in the source point cloud P, find the nearest point in the target point cloud Q as the corresponding point, which is achieved by calculating the Euclidean distance between all point pairs and taking the minimum value; Step 53: Use the least square method to solve the rotation matrix and the translation vector to minimize the objective function; Step 54: Repeat the above steps, and each iteration uses the transformation obtained in the previous iteration to update the position of the source point cloud P and re-search the corresponding points; the iteration continues until the preset stop condition is met, and the tire center position at time k is calculated ; In step 6, a characteristic evaluation function is designed based on the tire size characteristics, tire strength characteristics and motion characteristics, and the cost value is for: ； in, is the standard strength value of the target commercial vehicle tire, is the standard diameter value of the target commercial vehicle tire, yes k -1 The position of the tire center at the moment, is the position of the tire center at time k; I represents the point average intensity, Indicates tire size information.

2. The single-line laser radar tire detection method based on feature evaluation function according to claim 1, characterized in that: In the step 2, the original three-dimensional point cloud data is preprocessed, including removing outliers based on Gaussian filtering to obtain preprocessed three-dimensional point cloud data.

Images