An Surface Defect Extraction of Component Scratches Algorithm Based on Gaussian Filter

Abstract

Scratch defects on the surface of many tiny components are difficult to be solved by machine vision technology, and the algorithm needs to be constantly evolved. Therefore, based on image processing technology, a method of extracting scratch defects on the surface of components based on Gaussian filtering in frequency domain filtering is proposed. Firstly, in the early image preprocessing, the Gaussian filter is used to filter the image, and then the Hough transform algorithm is used to represent the contour, and the image morphology is analyzed. Finally, the Gabor filter is used to extract the scratch defects.