The Image Edge Detection Algorithm Research Of Black Piece Defect Inspecting System Based On Machine Vision

With the rapid development of modern industry and advanced manufacturing technologies, the product processing accuracy of continuous improvement, based on machine vision inspection system increased attention for the characteristics of non-contact, full field of view and high precision. This paper studies the edge detection based on machine vision technology in the Black piece defect detection application, papers around the image edge extraction technology to do the following work:First, discussed the Black piece machine vision technology in the detection of defects in the feasibility, necessity, and its theoretical and practical value. Illustrated the edge extraction technology in the Black piece based on machine vision inspection system in the need for detailed study of edge detection technology in the domestic and international dynamics. Described the definition and classification of the edge, and edge extraction Categories.Second,described the classical edge detection algorithm, and the algorithm is applied to the Black piece image edge extraction, given the strengths and weaknesses of various algorithms.Third, the studies based on mathematical morphology edge detection methods, detailed description of the basic theory of mathematical morphology which will be the value from the two morphological extended to gray-scale morphology. Use a multi-scale mathematical morphology edge detection algorithms to the Black piece edge detection, and conduct the related experiments, the experimental results and analysis.Fourth, describing the commonly used sub-pixel edge detection algorithms, such as curve fitting, interpolation method and moment method. On the basis of Zernike moments operator have the nature of orthogonal and is not sensitive to noise. But the algorithm for computing need a longer time, can not meet the Black piece online detection of real-time. This paper presents an image segmentation and improved combination of Zernike moments to carry out sub-pixel edge detection algorithm. Experimental results show that reduce the running time of the algorithm.