Research And Engineering Application Of Batch Parts Automatic Detection System Based On Machine Vision

Precision measurement and detection technology is the foundation of industrial develop-ment, the measurement and detection efficiency a certain extent determines the manufacturing industry development level. Image measuring system based on machine vision based on mo dern optics, optoelectronics,computer graphics, digital signal processing, automatic control a nd integration of science and technology for the non-contact measurement technology, is be coming a kind of efficiency measurement and detection and the key technology to guarantee the product quality, occupy important position in industrial development.This combination of Shanghai city science and technology project (09DZ1122600) the follow-up research and development of "support product digital manufacturing of research and development of high precision of image measurement system", the systematically, in-depth research of volume parts automatic measurement and detection base on machine vision has carried on, and had developed a batch parts automatic detection system based on machine vision. The main contents and achievements of this paper are as follows:(1)Extraction of digital image feature,such as straight line circular extraction base on the classic algorithm of Hough transform, feature point extraction algorithm Harris corner detection, and summarized in the experiment, the improved algorithm to extract the linear, circular, corner, automatic extraction module is developed at last part image feature.(2)In the aspect of image acquire of out of field of camera,the algorithm image mosaic that based on motion positioning control image mosaic are combined to form an adaptive algorithm for image mosaic:rich features parts using template matching algorithm for stitching; while part features rare, first of all, positioning error calibration for the work table, then, by using the method of motion positioning control to stitching, at the same time compensate the positioning error to stitching parameter, finally, to obtain adaptive panoramic image stitching.(3) In order to improve the measurement and detection efficiency of batch parts of image, this paper presents an automatic measuring and testing methods, through the machine teaching method for template production, design templates and characteristics, including measurement of characteristic value, and then through the image matching algorithm for pose detection and correction, the same batch of parts finally, through the automatic measurement and detection template to realize the same batch of parts.(4) The whole system improve and realization of engineering application.