Study On Image Processing Algorithm For Work Piece Testing

In order to measure the size and appearance of work piece quickly and accurately, an application system which combined the optics image formation technology with computer image processing technology was designed. Compared with the traditional measurement method, the automatic detection system of which belongs to untouched measurement, has high measurement precision, high sampling speed and is steady and reliable.In the paper, the measurement principle of the automatic work piece detection system was introduced, overall design, the composition of the hardware and software design were explained briefly, the influence of each component on systematic precision was analyzed, various kinds of methods of image pretreatment that the paper used were introduced.In the paper, a kind of improved Hough transformation image detection algorithm to realize the abstraction appearance outline of work piece, and the work piece's geometry size and conical degree were calculated by prior mathematics models . The algorithm has high efficiency at edge curve, and is less affected by noise and curve disconnection. The system proceeded the texture analysis with gray level cooccurrence matrix and statistical method, calculated and statistics the surface texture characteristic value of the work piece during the abstraction of goal work piece image, realized the appearance detection of the work piece surface by the parameter comparison, and finished the selection between eligible products and uneligibilie products after analyzing synthetically.