Defect Detection of Industrial Products based on Improved Hough Transform

In the industrial manufacturing process, defects of industrial products may inevitably occur. In order to detect the defects of industrial products with central symmetry distribution rules, this paper analyzes the limitations of traditional detection methods and proposes an improved detection method based on Hough transform. The method starts from the central area of the industrial product and performs directional clustering towards the direction in which the detection target is located: Firstly, obtain contour of the region of interest and the center point. Secondly, use Hough transform on the points on the contour of the Region Of Interest(ROI). Voting is performed according to the constraint rule that only passes through the direction of central area. The peak corresponding to the detected object is obtained in the voting space, and the defect of the industrial product component is obtained by estimating the peak position. Experimental results show that our algorithm has strong anti-interference ability and can solve the undetectable problems caused by the similarity between the detection target and the background. It can meet the requirements of certain types of industrial production and has significant robustness compared with the traditional Hough method.