Design and implementation of defect detection system for inner wall of bellows

Abstract Address the problems of low detection efficiency, high labor intensity and inconsistent detection rules of inner wall defects in the current bellows production process, a bellows inner wall defect detection system was designed and developed. According to the characteristics of small inner diameter and uneven inner wall of bellows, a side-view endoscope was adopted for image acquisition, and an execution mechanism and action process were designed to achieve comprehensive image acquisition of the inner wall of bellows. Deep learning algorithms were used to detect defects, and an upper computer software was designed and developed for on-site personnel to operate, view results and conduct report analysis. Experimental verification shows that the system can effectively identify different types of defects and effectively improve production efficiency and reduce labor intensity.