The Research On Automatic Workpiece Sorting System Based On Machine Vision

Sorting of work piece is a key step in the industrialization production, as sortingefficiency directly affects the efficiency of enterprises. Combined with the project, theplan of innovation activities in Universities of Shanghai,《multi axis automaticworkpiece sorting system based on machine vision》 against project number:(20131006), the key technology of computer vision measurement system is furtherstudied in this paper. Machine teaching method is used to create a feature library ofthe parts to be processed. Features of parts image are obtained by ImagePreprocessing, Threshold Segmentation, Edge and Contour Extraction algorithms, andcompared with standard templates, then the eligibility of tested parts are determined.To solve the position and angle deviations of parts, the edge detection and contourtracing algorithms are proposed to determine the center position and the rotation angleof the parts, and the similarity measurement is based on the template parts todetermine the characteristics of the parts,to provide decision support for automaticsorting. Effectiveness is certificated by such experiment.The main contents and results in this paper are as follows:(1) In the aspect of image processing algorithms, research key algorithms ofimage processing, such as image filtering, binarization, pretreatment, imagesegmentation, edge detection and contour tracing algorithm to get the center of figureand rotation angle of figure. This algorithm solves the problem of parts which may beplaced at random, and can be applied to the automatic sorting system. (2) In image recognition of machine vision, as standard workpieces templateswere pre-established, the work piece to be detected by travel and rotation transformcomparing with the standard template, can then be determined whether they aresimilar parts or not.(3) In the design of machine vision system hardware platform, the compositionof the system is introduced, besides combining with the necessity analysis ofindustrial field quality control, industrial CCD camera, image acquisition card, motorand gas claw hardware selection, a design of multi-channel light source structure, anda light source quality verification system is provided, thus certifies its reliability byexperiment.(4) In the aspect of software design of workpiece sorting system, the softwaredevelopment environment, as well as the system work flow are introduced, and thealgorithm is verified by MATLAB.