Recognition and Extraction Algorithm Design for Defect Characteristics of Armor-plate Flaw Detection Image

Aiming at the detection image of strip steel which rolled on the 15 mm plate production line in a steel plant, the defect characteristics recognition and extraction algorithm had been analyzed and designed, based on the computer image processing and pattern recognition theory. And the corresponding defect characteristics recognition and processing program had been programmed by VC++6.0 computer language. In the paper, the 8 direction pixel gray value search algorithm had been compiled based on the computer image color grading theory firstly, then to extract every gray level pixel information of the armor-plate detection image, and to carry out the corresponding every gray level pixel distribution probability statistic. Based on the statistical results, the two-dimension histogram Fish evaluation function algorithm for the armor-plate CCD image processing had been designed, and the result of practical application shows that the defect characteristics recognition system which programmed based on the algorithm ahead can accurately recognize and extract the defect characteristics data from the armor-plate rolled detection image, and can effectively satisfy the industrial production requirement of plate rolled.