3D point cloud defect detection system for tire tread quality inspection

Abstract

The invention discloses a 3D point cloud defect detection system for tire tread quality inspection, and relates to the technical field of industrial automatic detection, and the 3D point cloud defect detection system comprises a multi-modal data fusion module which combines a two-dimensional image and three-dimensional point cloud data and carries out multi-modal data registration through an adaptive optimization algorithm; the defect identification module is used for positioning abnormal characteristics of the surface of the tire through a residual analysis method and marking a defect area according to a set threshold value; the automatic control module is integrated with a tire production line, controls the tire to be detected through the grabbing device and the rotating device, and facilitates collection of 3D point cloud data of the tire; and the data visualization module is used for carrying out data processing operation and visually displaying the position and the type of the tire defect through a three-dimensional model. Therefore, by adopting the 3D point cloud defect detection system for tire tread quality inspection, the tire tread defects can be detected in real time, the detection precision and efficiency are effectively improved, and the 3D point cloud defect detection system has very high industrial practicability and popularization value.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial automation detection, and in particular to a 3D point cloud defect detection system for tire tread quality inspection. Background Art

[0002] With the development of the automobile industry, tires, as key components of automobiles, have a quality that is directly related to the safety and durability of vehicles. Therefore, in the tire production process, tread quality inspection becomes particularly important.

[0003] At present, tire defect detection mainly relies on manual inspection and two-dimensional image analysis. However, traditional manual inspection methods are inefficient and easily affected by the operator's subjective judgment, resulting in missed or false detection problems. At the same time, two-dimensional image detection technology cannot fully capture the three-dimensional features of the tire tread, especially in the case of complex tread patterns and non-rigid deformation of tire rubber materials. The plane information of the two-dimensional image is difficult to reflect the tiny height differences on the tread, which greatly limits the accuracy of defect detection.

[0004] In recent years, with the development of 3D sensing technology, 3D point cloud technology has been widely used in the field of industrial inspection. Through 3D laser scanners or other 3D sensors, three-dimensional data of the surface of an object can be obtained and more accurate geometric analysis can be performed. 3D point cloud technology has demonstrated its advantages in complex surface detection in many fields such as aviation, automobile manufacturing, and construction. However, 3D technology applied to tire inspection still faces some challenges, such as how to deal with non-rigid deformation of rubber materials, complex tread pattern structures, and interference from external noise. Therefore, a multimodal detection technology that can effectively fuse 2D and 3D data is needed, combined with accurate point cloud registration and defect recognition algorithms, to achieve efficient and reliable tire tread defect detection. Summary of the invention

[0005] The purpose of the present invention is to provide a 3D point cloud defect detection system for tire tread quality inspection, which can solve the problem of insufficient accuracy of existing 2D detection and overcome the difficulties of non-rigid deformation and noise processing in 3D detection, thereby improving the reliability and accuracy of detection.

[0006] To achieve the above object, the present invention provides a 3D point cloud defect detection system for tire tread quality inspection, comprising: The 3D point cloud data acquisition module is equipped with a line laser sensor and is arranged on the tire production line to obtain the 3D point cloud data of the tire tread and sidewall; Multimodal data fusion module, combining 2D images and 3D point cloud data, and performing multimodal data registration through adaptive optimization algorithm; The defect recognition module locates the abnormal features on the tire surface through the residual analysis method and marks the defect areas according to the set threshold; The automatic control module is integrated with the tire production line and controls the tire to be detected through the grasping device and the rotating device, facilitating the acquisition of the 3D point cloud data of the tire; The data visualization module is used for data processing operations and intuitively displays the position and type of tire defects through a 3D model.

[0007] Preferably, the line laser sensors are arranged at different angles.

[0008] Preferably, multi-modal data registration includes: First, a two-dimensional image of the tire surface is collected by a camera or a two-dimensional imaging device, and the preliminary spatial position information is provided by the two-dimensional image and key feature points are extracted to initially obtain the preliminary pose information of the tire in the three-dimensional space; Then, using the obtained preliminary pose information, the closest point pairs between the target point cloud and the source point cloud are found through the iterative closest point algorithm, the rigid transformation matrix is calculated, and the position information of the source point cloud is updated until the iteration ends, completing the detailed registration of the 3D point cloud data.

[0009] Preferably, the residual analysis method includes: First, a three-dimensional point cloud model of the standard tread is generated; Secondly, the difference between the point cloud data of the tire to be detected and the standard model is calculated to generate a residual value; Next, the residual is analyzed according to the characteristics of the normal tread pattern and the actual defect, and a suitable threshold is set; Then, the areas where the residual values exceed the preset threshold are marked as defect areas, and a corresponding inspection report is generated.

[0010] Preferably, the rotating device drives the tire to rotate using a servo motor.

[0011] Therefore, the present invention adopts the above 3D point cloud defect detection system for tire tread quality inspection, having the following technical effects: (1) It has data processing and visualization functions, can intuitively display the specific position and type of defects in the form of a 3D model and image, helps operators quickly evaluate and locate problems, and provides an effective quality control and decision-making support tool for the factory management.

[0012] (2) The visualization function also supports generating inspection reports for production management personnel to conduct further data analysis and production process optimization, enhancing the intuitiveness of the inspection results.

[0013] (3) Combining multimodal data registration and residual analysis algorithms, it can effectively solve the impact of subtle burrs on the rubber surface, surface deformation and environmental noise on the detection accuracy, and maintain efficient and stable detection results in complex industrial environments.

[0014] (4) Realize automated testing operations, reduce reliance on manual testing, and reduce production costs. At the same time, it can provide real-time feedback on test results, have powerful data processing and visualization capabilities, improve management efficiency, and provide an effective reference for long-term data analysis.

[0015] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A schematic diagram of visualization of 3D point cloud data in an embodiment of a 3D point cloud defect detection system for tire tread quality inspection; Figure 2 A schematic diagram of a multimodal data registration process in an embodiment of a 3D point cloud defect detection system for tire tread quality inspection; Figure 3 It is a schematic diagram of the residual analysis and defect identification process in an embodiment of a 3D point cloud defect detection system for tire tread quality inspection; Figure 4 The present invention is a schematic diagram of defect detection results in an embodiment of a 3D point cloud defect detection system for tire tread quality inspection, wherein a represents a first group of tires and b represents a second group of tires. DETAILED DESCRIPTION

[0017] The present invention can be explained in more detail by the following examples. The purpose of disclosing the present invention is to protect all changes and improvements within the scope of the present invention. The present invention is not limited to the following examples.

[0018] The current quality inspection of tire treads mainly relies on two-dimensional images or manual inspection, which shows great limitations when dealing with complex tread patterns and small height changes. Existing two-dimensional detection methods cannot capture the three-dimensional features of the surface, especially under the non-rigid deformation conditions of rubber materials, which often leads to high false detection and missed detection rates. In addition, although the existing three-dimensional detection methods can obtain geometric information on the tire surface, when processing tire treads, they are limited by the matching accuracy of the algorithm and its adaptability to non-rigid deformation. It is difficult to effectively remove noise while ensuring high precision, resulting in inaccurate detection results. To this end, the present invention provides a 3D point cloud defect detection system for tire tread quality inspection, comprising: (1)3D Point Cloud Data Acquisition Module: A high-precision line laser sensor is arranged on the tire production line to perform full-coverage scanning on the rotating tire, obtain 3D point cloud data of the tire tread and sidewall, and combine multi-angle scanning of multiple sensors to achieve blind-spot-free detection of the tire in all directions, and generate a 3D model to intuitively represent the details and surface features of the tread, such as Figure 1 shown.

[0019] At the same time, after the tire enters the detection area, the tire is fixed at a specified position by a grasping device, and the rotation of the tire is controlled by a servo motor to ensure stable rotation of the tire, effectively reducing data distortion caused by mechanical vibration and improving the accuracy of the collected data.

[0020] (2)Multi-modal Data Fusion Module: Combine two-dimensional images and three-dimensional point cloud data, make full use of the edge information of 2D images and the height information of 3D point clouds to ensure accurate alignment at complex patterns, thereby reducing errors, and realize the registration of multi-modal data through the Iterative Closest Point (ICP) algorithm to ensure that even when the rubber material undergoes non-rigid deformation, the data can still be accurately aligned.

[0021] As Figure 2 shown, the specific process of multi-modal data registration is as follows: First, collect two-dimensional images of the tire surface through a camera or two-dimensional imaging device, and these images can provide the basic position and shape features of the tire surface. Second, use the two-dimensional images to provide preliminary spatial position information and extract key feature points, such as the edges, patterns or marks of the tire. Then, through computer vision algorithms (such as feature matching or template matching), the pose (position and orientation) of the tire in three-dimensional space can be roughly estimated. Then, use the Iterative Closest Point (ICP) algorithm to perform fine registration on the three-dimensional point cloud data to ensure high-precision alignment even when the tire undergoes non-rigid deformation and has better adaptability.

[0022] Among them, the Iterative Closest Point (ICP) algorithm for detailed registration of three-dimensional point cloud data includes: First, high-precision point cloud data of the tire surface is obtained through a three-dimensional laser scanner or other three-dimensional sensors, and these data contain the three-dimensional coordinates of each position on the tire surface. Second, using the preliminary pose information obtained from the two-dimensional image, the collected three-dimensional point cloud data is roughly aligned. Although there may be a certain pose error between the point cloud and the feature points of the two-dimensional image at this time, there is already a general matching relationship. Next, the ICP algorithm is used to further refine the result of point cloud registration. By continuously iterating, the closest point pairs between the target point cloud and the source point cloud are found, and a rigid transformation matrix (including a rotation matrix and a translation vector) is calculated to align the source point cloud as closely as possible with the target point cloud. Then, applying the calculated transformation matrix, the position of the source point cloud is updated and gradually approaches the shape of the target point cloud. The registration ends when the registration error is less than the set threshold or the predetermined maximum number of iterations is reached.

[0023] This adaptive optimization algorithm can dynamically adjust the registration parameters and adaptively adjust the registration strategy between the point cloud and the image according to the changes in the tire surface characteristics. Especially when facing the non-rigid deformation of rubber materials, it can achieve higher adaptability and accuracy.

[0024] (3) Defect recognition module: According to the collected point cloud data of the tire surface, through the residual analysis algorithm, the difference between the detection area and the standard tire surface model is calculated, and by calculating the residual values, surface defects such as protrusions, depressions, and cracks on the tire surface are identified.

[0025] As Figure 3 shown, the specific process of defect recognition includes: First, a three-dimensional point cloud model of the standard tire surface is generated as a reference; second, the difference between the point cloud data of the tire surface to be detected and the standard model is calculated to generate residual values; then, these residuals are analyzed to distinguish normal tire tread patterns and actual existing defects (such as protrusions, depressions, etc.); then, the areas where the residual values exceed the preset threshold are marked as defect areas, and corresponding detection reports are generated.

[0026] Compared with traditional technologies, this residual analysis method can not only accurately distinguish normal patterns and surface defects, but also dynamically adjust the sensitivity according to the actual application scenario to adapt to the patterns and structures of different tires. Compared with the existing technologies, the method of this embodiment improves the recognition rate of micro defects, significantly reduces the phenomena of false detection and missed detection, reduces the defective rate on the production line, improves the overall quality and reliability of the product, and avoids potential safety hazards caused by tire defects. This technical advantage directly improves the production efficiency of the enterprise, reduces the after-sales risk, and enhances the competitiveness of the product in the market.

[0027] (4) Automatic control module: Integrated with the tire production line, it can achieve automated operations on the tire production line. Among them, the automatic control system includes a gripping device and a rotating device to ensure that the tire is in an ideal position during data collection. The equipment can automatically complete the detection task without manual intervention, which not only improves the detection speed and accuracy but also reduces the dependence on manual inspection, further reducing production costs.

[0028] The detection system can be seamlessly integrated with the tire production line to achieve fully automated detection, which not only reduces manual intervention but also improves the detection efficiency and real-time nature of the detection results, significantly enhancing the production capacity of the production line. The tire is automatically transported to the detection system through the production line. At this time, the automatic control module starts to work, gripping and rotating the tire. Meanwhile, the laser sensor scans the tread to collect three-dimensional point cloud data.

[0029] After the data collection is completed, the system will automatically process the data, perform multi-modal registration and residual analysis, and finally output the detection results, which are fed back to the production control system to automatically determine whether the tire is a defective product.

[0030] (5) Data visualization module: Used to visually display the specific location and type of defects in the form of three-dimensional models and images, helping operators quickly evaluate and locate problems; it also supports generating detection reports for production management personnel to conduct further data analysis and optimize the production process, providing an effective quality control and decision-making support tool for the factory management. As Figure 4 shown, the detection results of two groups of tire surface defects are presented. The defects on each tread can be accurately detected, and the specific location and detection accuracy of the detected tire surface defects, including protrusions, depressions, etc., can be visually represented through the visualization module.

[0031] Therefore, by adopting the above 3D point cloud defect detection system for tire tread quality inspection, the present invention can effectively improve the accuracy and efficiency of tire tread defect detection, and has strong industrial practicability and popularization value.

[0032] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that they can still modify or equivalently replace the technical solutions of the present invention, and these modifications or equivalent replacements cannot make the modified technical solutions deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A 3D point cloud defect detection system for tire tread quality inspection, characterized in that: include: The 3D point cloud data acquisition module is equipped with a line laser sensor and is arranged on the tire production line to obtain the 3D point cloud data of the tire tread and sidewall; Multimodal data fusion module, combining 2D images and 3D point cloud data, and performing multimodal data registration through adaptive optimization algorithm; The defect recognition module locates abnormal features on the tire surface through residual analysis and marks defective areas according to the set threshold; The automatic control module is integrated with the tire production line to control the tire to be tested through the gripping device and the rotating device, which facilitates the collection of tire 3D point cloud data; The data visualization module is used for data processing operations and intuitively displays the location and type of tire defects through a three-dimensional model.

2. A 3D point cloud defect detection system for tire tread quality inspection according to claim 1, characterized in that: Line laser sensors are arranged at different angles.

3. The 3D point cloud defect detection system for tire tread quality inspection according to claim 1, characterized in that: Multimodal data registration, including: First, a two-dimensional image of the tire surface is collected by a camera or a two-dimensional imaging device, and the two-dimensional image is used to provide preliminary spatial position information and extract key feature points to obtain preliminary posture information of the tire in three-dimensional space. Then, using the obtained preliminary pose information, the iterative closest point algorithm is used to find the nearest point pair between the target point cloud and the source point cloud, calculate the rigid transformation matrix, and update the source point cloud position information until the iteration is completed to complete the detailed alignment of the three-dimensional point cloud data.

4. The 3D point cloud defect detection system for tire tread quality inspection according to claim 1, characterized in that: Residual analysis methods, including: First, a 3D point cloud model of the standard tread is generated; Secondly, the difference between the tread point cloud data to be inspected and the standard model is calculated to generate a residual value; Next, the residuals are analyzed based on the characteristics of normal tread patterns and actual defects, and appropriate thresholds are set; Then, the area where the residual value exceeds the preset threshold is marked as a defective area, and a corresponding inspection report is generated.

5. The 3D point cloud defect detection system for tire tread quality inspection according to claim 1, characterized in that: The rotating device uses a servo motor to drive the tire to rotate.

Images