Detection method of micro and macro scale defects in silicon carbide castings

Abstract

Aiming at challenges such as low efficiency, high missing rate, difficulty in identifying contour defects, and difficulty in extracting tiny defects, a defect detection method for extracting micro and macro scale defects is proposed in this paper. After preprocessing the image, contour detection is performed to identify the contours. Subsequently, a contour complementation algorithm is employed to complement the unclosed contours. Finally, the detection of micro scale defects is conducted based on the grayscale variation of the center of the micro scale defects. The experimental results show that compared with the traditional method, the proposed algorithm can accurately detect the bubble defects of different scales in silicon carbide castings and can identify the complex defects better.