Automobile tire defect automatic detection equipment and system

Abstract

The invention discloses an automobile tire defect automatic detection device and system, and belongs to the field of automobile tire processing, through highly integrated modular design, the automobile tire defect automatic detection device comprises a multifunctional image recognition module, a depth measurement and 3D imaging module, an internal structure and air tightness detection module and an analysis and decision module, comprehensive detection of the surface, the geometric structure, the internal structure and the air tightness of the tire is realized, and efficient, accurate and intelligent tire defect detection is realized through cooperative work of each module and equipment.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of automobile tire processing, and specifically relates to an automobile tire defect automatic detection device and system. BACKGROUND

[0002] Automobile tires, as key components of vehicles, their quality and performance are directly related to driving safety and comfort. Therefore, tire detection and quality control are of great significance in the process of automobile manufacturing and maintenance. At present, the commonly used automobile tire detection methods mainly include the following:

[0003] Visual detection (manual visual detection and machine visual detection), mainly used to identify the obvious defects such as wear, cracks, bubbles on the surface of the tire;

[0004] Mechanical detection (pressure, hardness, ultrasonic detection, etc.), mainly used to identify the integrity and air tightness of the tire inside;

[0005] However, manual visual detection relies on the subjective judgment of the inspector, and is easily affected by factors such as fatigue and lack of experience, resulting in low reliability and consistency of the detection results. Machine visual detection, although it improves the consistency of detection, still has limited recognition accuracy in identifying small defects or defects in complex backgrounds, especially in conditions of large changes in light conditions, and still relies on manual operation, with low automation level, which is difficult to meet the needs of large-scale production and rapid detection, resulting in low detection efficiency;

[0006] For mechanical detection, high-end detection equipment such as high-resolution multispectral cameras, laser scanners, X-ray devices and high-performance ultrasonic sensors are expensive, increasing the overall cost of the tire detection system. At the same time, professional technicians are needed to maintain and operate these devices, further increasing the operating cost.

[0007] In summary, the existing automobile tire detection methods have many shortcomings in detection accuracy, coverage, automation level, real-time performance, cost control and data processing, etc. These defects limit the efficiency and accuracy of tire quality detection, making it difficult to meet the needs of modern automobile manufacturing and maintenance for high-quality and efficient tire detection.

[0008] Therefore, an integrated, intelligent, multi-modal data fusion detection system is urgently needed to make up for the shortcomings of existing methods and achieve the comprehensiveness, accuracy and efficiency of tire detection SUMMARY

[0009] In view of the above defects, the present application provides an automatic tire defect detection system for automobiles, which realizes high integration and collaborative work through different division of labor among multiple modules and unified control to complete multi-dimensional detection tasks such as tire surface, geometric structure, internal structure and air tightness, and specifically includes the following modules:

[0010] A multifunctional image recognition module for detecting tire patterns and tire sidewall cracks, including an image acquisition unit for acquiring tire surface and sidewall images, a first edge computing unit for pre-processing the acquired images to improve image quality, a multi-task defect recognition module for detecting patterns and cracks in the pre-processed images;

[0011] A depth measurement and 3D imaging module for detecting tire geometric structure defects and generating a geometric model, including a 3D model data acquisition unit for acquiring 3D data of the tire, a second edge computing unit for pre-processing the acquired 3D model data, a geometric defect detection unit for analyzing the tire shape based on the pre-processed 3D model data, and a data correlation unit for spatially aligning the 3D model data with the images pre-processed by the first edge computing unit;

[0012] An internal structure and air tightness detection module for positioning and evaluating tire internal defects and detecting air tightness defects, including an ultrasonic signal unit for capturing tire internal information, a third edge computing unit for extracting signal features of the internal information, a defect positioning unit for calculating defect positions and depths based on the ultrasonic signals, and a correlation and verification unit for correlating the calculated internal defect data with the data from the first and second edge computing units;

[0013] An analysis and decision module for analyzing the data output by the multifunctional image recognition module, the depth measurement and 3D imaging module, and the internal structure and air tightness detection module and generating a detection report and repair recommendations.

[0014] Further, the first edge computing unit has the application of denoising, contrast enhancement and edge detection technology to improve image quality.

[0015] Further, the second edge computing unit has cloud filtering, noise removal and data alignment technology to generate a complete three-dimensional model.

[0016] Further, the third edge computing unit extracts signal features through Fourier transform technology and identifies internal defect types in real time through a machine learning model.

[0017] Further, the detection report generated by the analysis and decision module includes classified defect types and priority ranking:

[0018] Classification defect types include:

[0019] Good: no obvious defects;

[0020] Surface defects: pattern wear, irregular scratches;

[0021] Structural defects: geometric asymmetry, sidewall bulging;

[0022] Internal defects: wire breakage, air tightness problems;

[0023] Comprehensive defects: multiple defects exist at the same time;

[0024] Priority ranking includes:

[0025] High priority: internal contact breakage, serious air tightness problems;

[0026] Medium priority: sidewall cracks, partial bulging;

[0027] Low priority: slight pattern wear, surface scratches.

[0028] The present application also discloses a device applied to the above-mentioned automatic detection system of automobile tire defects, including the following devices:

[0029] Multispectral high-resolution camera and light source regulator applied to multifunctional image and recognition module;

[0030] High-precision laser scanner and structured light sensor applied to depth measurement and 3D imaging module;

[0031] Several ultrasonic sensors and acoustic analyzers distributed around the tire applied to internal structure and air tightness detection module;

[0032] Processing terminal applied to analysis and decision module, including processor, data storage unit, data transmission unit and touch screen for interactive interface operation;

[0033] And automatic rotating platform for fixing the tire.

[0034] Compared with the prior art, the present application has the following beneficial effects:

[0035] By reducing the number and cost of different devices used for different detection requirements of existing tires through the overall system and device, the flexibility and scalability of the system are improved;

[0036] Comprehensive utilization of image, 3D geometry and ultrasonic data improves the comprehensiveness and accuracy of detection;

[0037] Utilizing machine learning and edge computing, automatic defect recognition, real-time report generation and intelligent decision-making are realized;

[0038] Local data processing is realized through edge computing, and overall response speed of the system is improved;

[0039] Through the highly integrated modular design, the overall detection of the tire surface, geometric structure, internal structure and air tightness is realized, through the collaborative work of the multifunctional image recognition module, the depth measurement and 3D imaging module, the internal structure and air tightness detection module and the analysis and decision module, the efficient, accurate and intelligent tire defect detection is realized. BRIEF DESCRIPTION OF DRAWINGS

[0040] Figure 1 The working flowchart of the analysis and decision module in the application is shown in the figure. DETAILED DESCRIPTION

[0041] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only a part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the application.

[0042] EMBODIMENT

[0043] The embodiment provides an automatic tire defect detection system for a vehicle, which comprises the following modules:

[0044] The multifunctional image recognition module is used for detecting tire patterns and tire sidewall cracks, and simultaneously collecting high-resolution and multi-spectrum images, and specifically comprises:

[0045] 1) an image collection unit for collecting tire surface and sidewall images;

[0046] 2) a first edge computing unit for pre-processing the collected images to improve image quality, by applying denoising, contrast enhancement, edge detection and other techniques to improve image quality;

[0047] 3) a multi-task defect recognition module for detecting patterns and cracks in the pre-processed images, which utilizes a pre-trained lightweight deep learning model such as a MobileNet model to perform real-time pattern wear and crack detection on the edge computing unit. For pattern depth detection, a pattern depth distribution map is generated to identify worn areas, and for sidewall crack detection, the crack position and its geometric features (length, width) are labeled;

[0048] The depth measurement and 3D imaging module is used for detecting tire geometric structure defects (such as sidewall bulging, asymmetric wear, etc.) and generating high-precision three-dimensional geometric models, and specifically comprises:

[0049] 1) 3D model data acquisition unit for 3D data acquisition (three-dimensional point cloud data) of the tire;

[0050] 2) Second edge computing unit for preprocessing the acquired 3D model data, for point cloud filtering, noise removal, data alignment, to generate a complete three-dimensional model;

[0051] 3) Geometric defect detection unit for analyzing the tire shape of the preprocessed 3D model data, identifying tire shape abnormalities by using real-time geometric analysis algorithms, including sidewall bulges, asymmetric wear, etc., identifying local swelling areas that exceed the set threshold, i.e. determining as bulges, analyzing the spatial distribution of pattern depth, identifying uneven wear problems;

[0052] 4) Data association unit for spatial alignment of 3D model data and image preprocessed by the first edge computing unit to enhance the recognition accuracy of geometric defects;

[0053] Internal structure and air tightness detection module for detecting internal air tightness defects, steel wire breakage, and cord layer damage, etc. and achieving precise positioning and evaluation of internal defects through ultrasonic and acoustic signals, specifically including:

[0054] 1) Ultrasonic signal unit for controlling the sending and receiving of reflected signals to capture internal structure information of the tire;

[0055] 2) Third edge computing unit for extracting signal features of internal information, specifically using advanced signal processing techniques such as Fourier transform and wavelet transform to extract signal features, while using machine learning models (such as SVM, random forest) to identify internal defect types (air tightness problems, steel wire breakage, cord layer damage) in real time on the edge computing unit;

[0056] 3) Defect positioning unit for calculating the specific location and depth of defects based on the time difference and intensity change of ultrasonic signals;

[0057] 4) Association and verification unit for associating the calculated internal defect data with the data of the first and second edge computing units to verify the coherence and accuracy of the defects;

[0058] As shown in Figure 1 the analysis and decision module analyzes the data output by the multifunctional image recognition module, the depth measurement and 3D modeling module, and the internal structure and air tightness detection module, and generates a detection report and repair recommendations, with the specific workflow being:

[0059] S1, collect the pattern and crack image data of the multifunctional image recognition module, the three-dimensional geometric model data of the depth measurement and 3D imaging module, and the internal structure data of the internal structure and air tightness detection module, and then integrate the multi-modal data by applying data fusion technology (such as feature-level fusion and decision-level fusion) to form a unified data set;

[0060] S2, analyze the data set by using a deep neural network model (other machine learning models such as multilayer perceptron and ensemble learning algorithm can also be used), and obtain the following defect types:

[0061] Good product: no obvious defects;

[0062] Surface defects: pattern wear, irregular scratches;

[0063] Structural defects: geometric asymmetry, sidewall bulge;

[0064] Internal defects: wire breakage, air tightness problem;

[0065] Comprehensive defects: multiple defects exist at the same time;

[0066] S3, automatically generate repair priorities according to the types and severity of defects, and the priorities are:

[0067] High priority: internal contact breakage, serious air tightness problem;

[0068] Medium priority: sidewall crack, partial bulge;

[0069] Low priority: slight pattern wear, surface scratch;

[0070] S4, display the detection results through a graphical user interface, intuitively mark the defect location and type, and automatically generate a detailed detection report including defect description, location, severity and repair suggestion (for high priority defects, the system automatically triggers an alarm to notify maintenance personnel for emergency handling).

[0071] It should be noted that in order to enable the above-mentioned automobile tire defect automatic detection system to normally perform detection, hardware support is also required, which specifically includes the following devices:

[0072] When cooperating with the multifunctional image recognition module, a multispectral high-resolution camera and a light source regulator are required. The multispectral high-resolution camera (Teledyne DALSA) supports multi-band imaging such as visible light, near-infrared light and ultraviolet light, while the light source regulator includes LED light source, laser illumination, etc., providing illumination at different angles and wavelengths;

[0073] When cooperating with the depth measurement and 3D imaging module, a high-precision laser scanner and a structured light sensor are needed. The high-precision laser scanner (such as Leica Geosystems) can quickly obtain high-density point cloud data, and the structured light sensor (Intel RealSense) is used to supplement the depth information.

[0074] When cooperating with the internal structure and air tightness detection module, a multi-angle ultrasonic sensor and an acoustic analyzer are needed. The multi-angle ultrasonic sensor is arranged in an array at different positions of the tire to realize multi-directional signal transmission and reception, and the acoustic analyzer (such as Olympus) is used for high-precision signal analysis.

[0075] The analysis and decision module needs a terminal server, including a GPU (for complex data processing and machine learning model running), a data storage unit (for centralized storage and management of detection data of each module), an algorithm library (pre-installed with multiple algorithms for data fusion and intelligent analysis), a data transmission unit (wireless data transmission module) for data transmission between each module and the terminal server, and a touch screen (for display and control of GUI).

[0076] In addition, an automatic rotating platform for fixing the tire is also needed, and the tire is limited by the fixing part.

[0077] The overall detection process of the system is as follows:

[0078] First, the tire (named T123) is fixed on the automatic rotating platform to ensure stable rotation, and then the detection parameters of each module (such as camera exposure time, laser scanning density, ultrasonic frequency, etc.) are adjusted manually according to the model and specifications of the tire. It needs to be noted that the tire model can also be identified through RFID or two-dimensional code technology, which will not be described here.

[0079] Then the multifunctional image recognition module starts working: the multispectral camera synchronously collects the surface and sidewall images of the tire T123 under visible light, near-infrared and ultraviolet spectra, the first edge computing unit performs image preprocessing, identifies pattern wear unevenness and micro-cracks, generates a pattern wear area annotation map (depth range 0.3-0.6mm), and generates sidewall crack position and length data (crack 1: position S1, length 8mm, width 0.2mm; crack 2: position S2, length 12mm, width 0.3mm).

[0080] Then the depth measurement and 3D imaging module starts working: the laser scanner and the structured light sensor are enabled, and the three-dimensional point cloud data of the tire T123 is collected, the second edge computing unit performs point cloud filtering, generates a three-dimensional model of the tire T123, and identifies sidewall bulges (volume 3mm 3) and asymmetric wear, the data correlation unit confirms that the bulge position is consistent with the crack position detected by the multifunctional image recognition module, indicating a structural problem;

[0081] Then the internal structure and air tightness detection module starts to work, the multi-angle ultrasonic sensor sends and receives signals, and the internal structure detection is performed. The third edge computing unit identifies that the steel wire is broken, the breakage position C1, and the length is 15 mm. The correlation and verification unit combines the data of the multifunctional image recognition module and the depth measurement and 3D imaging module, and confirms that the internal steel wire breakage causes the external crack and the structural bulge;

[0082] Finally, the analysis and decision module analyzes the data output by the multifunctional image recognition module, the depth measurement and 3D imaging module, and the internal structure and air tightness detection module, and generates a detection report and a repair suggestion. According to the above data, the following detection report and repair suggestion can be obtained:

[0083] Classification: unqualified;

[0084] Defect type: surface defect, structural defect, internal defect;

[0085] Repair suggestion: replace the tire.

[0086] It should be noted that the structure described in the present application can be implemented in many different forms, and is not limited to the embodiments described. Any equivalent transformation made by those skilled in the art using the contents of the present application specification and drawings, or direct or indirect application in other related technical fields, such as other article loading and unloading, is included within the scope of protection of the present application.

Claims

1. The automatic detection system for automobile tire defects is characterized by: Includes the following modules: A multifunctional image recognition module for detecting tire tread patterns and tire sidewall cracks, comprising an image acquisition unit for capturing images of the tire surface and sidewall, a first edge computing unit for preprocessing the acquired images to improve image quality, and a multi-task defect recognition module for detecting tread patterns and cracks in the preprocessed images; A depth measurement and 3D imaging module, configured to detect geometric structural defects in tires and generate a geometric model, includes a 3D model data acquisition unit for acquiring 3D data of the tire, a second edge computing unit for preprocessing the acquired 3D model data, a geometric defect detection unit for analyzing the tire shape based on the preprocessed 3D model data, and a data association unit for spatially aligning the 3D model data with the image preprocessed by the first edge computing unit; An internal structure and airtightness detection module, used to locate and evaluate internal defects in the tire, as well as detect airtightness defects, includes an ultrasonic signal unit for capturing internal tire information, a third edge computing unit for extracting signal features of the internal information, a defect location unit for calculating the defect position and depth based on the ultrasonic signal, and an association and verification unit for associating the calculated internal defect data with data from the first and second edge computing units; The analysis and decision-making module analyzes the data output by the multi-functional image recognition module, the depth measurement and 3D forming module, and the internal structure and airtightness detection module and generates a detection report and repair suggestions.

2. The automatic tire defect detection device and system according to claim 1, characterized in that: The first edge computing unit is capable of applying denoising, contrast enhancement and edge detection technologies.

3. The automatic tire defect detection device and system according to claim 1, characterized in that: The second edge computing unit is equipped with cloud filtering, noise removal, and data alignment technologies to generate a complete three-dimensional model.

4. The automatic tire defect detection device and system according to claim 1, characterized in that: The third edge computing unit extracts signal features through Fourier transform technology and identifies internal defect types in real time through a machine learning model.

5. The automatic tire defect detection device and system according to claim 1, characterized in that: The inspection report generated by the analysis and decision module includes classification of defect types and priority ranking: Classified defect types include: Good product: no obvious defects; Surface defects: pattern wear, irregular scratches; Structural defects: geometric asymmetry, sidewall bulge; Internal defects: steel wire breakage, airtightness problems; Comprehensive defects: multiple defects exist at the same time; Prioritization includes: High priority: internal connection broken, serious airtightness problem; Medium priority: side wall cracks, partial bulging; Low priority: Minor tread wear, surface scratches.

6. Automatic automobile tire defect detection equipment, applied to the automatic automobile tire defect detection system according to any one of claims 1 to 5, characterized in that: Includes the following equipment: Multispectral high-resolution cameras and light source regulators used in multifunctional image and recognition modules; High-precision laser scanners and structured light sensors used in depth measurement and 3D imaging modules; Several ultrasonic sensors and acoustic analyzers distributed around the tire used in the internal structure and airtightness detection module; The processing terminal used in the analysis and decision-making module includes a processor, a data storage unit, a data transmission unit, and a touch screen for interactive interface operations; and an automatic rotating platform for securing tires.

Images