Research On Key Image Processing Technique Of Tire Section

The contour dimension of tires'section affects their capability, life span, safety and comfort directly, which is the significant factor of tires'quality. The traditional ways to measure section contour of tires are too empirical and inaccurate to match the digital design. With the rapid development of the motor industry and the priority to the development of radial tires, the research on the visualization detection system of tires'section has important theoretical significance and engineering value in China. It is helpful to improve the design and the detection level of radial tires. Image mosaic and extraction of section contour are the key technologies of the testing system, which is directly related to the precision and the reconstruction of tires'section contour. Therefore, based on the object of the research, this paper conducts systematical research work on the image mosaic algorithm of tires'section, the extracting algorithm on the edge of apex and tire section contour.Research on image mosaic algorithm. According to the principle of cross entropy stemming from information theory, an image matching principle based on cross entropy is proposed and the corresponding matching algorithm is studied in this paper. Besides, the image matching principle based on phase correlation is applied to the matching of tires' section images and the corresponding matching algorithm is studied as well. Finally, an image fusion algorithm was studied to achieve the fusion as well as the edge protection.Research on the edge extracting of apex, In the edge extracting of apex, the classic edge extraction methods are analyzed. Besides, According to the low constrast and gray gradients on the edge of apex, the methods to test the edge of apex and to preprocess the section images based on binary imaging morphology are proposed.Research on the extraction of tires'contour. Based on the characteristics of the tires' section images, some threshold segmentation methods are applied to the tires'section images, their corresponding experimental results are analyzed as well. According to the low constrast of tires'contour as well as the uneven target and background, the threshold segmentation algorithm based on gray imaging morphology is proposed, the effectiveness is verified by experiments.At last, the achievement of the research is summarized and the further research direction is guided as well.