Research And Implementation Of Automobile Tire Tread Defect Detection System

With the increase of automobile ownership,traffic accidents become more and more frequent,among which the death rate of tire blowout is extremely high.The defect of tire is the main cause of tire blowout.How to accurately and efficiently detect the defect of tire is a key research direction for the tire inspection industry.The main purpose of the paper is to study and implement a set of automobile tire tread defect detection system.The system is based on machine vision to carry out intelligent detection of tread defects,which not only ensures the accuracy and objectivity of detection results,but also improves the detection efficiency.The research provides a new idea and method for intelligent tire detection.The main contents of the paper include the study of automobile tire tread image extraction,the study of tread defect detection algorithm,the implementation and testing of tread defect detection system.In order to avoid the influence of the background in the image on the detection process and the detection result,the algorithm of tread image extraction is studied.The internal parameter matrix and distortion matrix were obtained by camera calibration.Correct the distortion of the tire image.Gamma transform was used to enhance the image.Canny operator is used for edge detection.Flood filling method is used to fill the area.Morphological processing is carried out by using closed operation.Extract the maximum connected domain.Logical "AND" operation is performed for the image of the maximum connected domain and the tire image.Finally,the image extraction of tread is realized.The test results show that the algorithm can effectively extract the tire tread from the tire image.The algorithm of tread defect detection was studied deeply,and the corresponding algorithm was built according to different types of tread defects to realize the detection.In view of the detection of tread wear,the Gray-Level Co-occurrence Matrix of tread image was calculated,and four characteristic values were analyzed:Energy,Entropy,Contrast and inverse different Moment.The degree of tread wear was determined comprehensively according to the results of characteristic values;In view of the detection of missing tread block,the tread image is divided according to the extreme point of 0 degree Radon transform.The position of missing blocks is determined according to the results of 90 degree Radon transform of segmented images;In view of the detection of tread crack,the image characteristics of tread cracks were analyzed,and the maximum entropy threshold segmentation algorithm was used to segment the target region of the image.The connected domain algorithm and the measurement of target circularity were used to identify the cracks.The crack size calculation method was developed,and the safety level was determined according to the crack width.All kinds of defect detection algorithms have been verified and tested,and the expected results have been achieved.The hardware test platform and software development platform are built according to the overall scheme of the system.The software design is carried out according to the algorithm flow of the research,and the programming and software debugging of the system are carried out.The system uses PyCharm integrated development environment,uses OpenCV computer vision library,uses Python to programming,realizes the acquisition,transmission,analysis and processing of tire image.The operation interface of the system is developed by using Qt Designer combined with Pyside2 interface library,and the detection results are displayed on the interface.Using MySQL database to store the detection results and other information,and finally realize the function of the system.The system integrates the functions of tire image acquisition,wireless transmission,analysis and processing,interface display,database storage and so on,and realizes the intelligent detection of tire tread defects.The validity and reliability of the system are verified by the test and result analysis of the system.The test results show that the accuracy of the system reaches the expected goal.