A Machine Vision System for Film Capacitor Defect Inspection

In order to solve several problems such as less automation, high labor intensity and inefficiency, which are brought by the traditional human inspection methods, we develop a machine vision system for film capacitor defect inspection in this paper. Specifically, we use high-definition CMOS industrial cameras, IPC board and STM32 controller to set up the hardware platform, and put forward suitable image processing algorithms for film capacitor appearance defect inspection, such as forward extraction, horizontal and vertical projection, shape detection and gradient detection. The proposed algorithms can well realize the detection of uneven appearance defect, glitch defect, red paint missing defect and protruding edge defect of film capacitors. The experimental results show that the proposed vision system has high accuracy and reliability, can meet the modern production requirements.