A robust vision inspection system for detecting surface defects of film capacitors

• We construct a robust vision inspection system to detect surface defects of film capacitors.
• We propose a effective vision-based defect detection approach based on Non-subsampled Contourlet Transform.
• We report the superior performance by the system compared to existing solutions.

This paper presents a robust vision inspection system for detecting the surface defects of film capacitors. In particular, we use a novel Non-subsampled Contourlet Transform (NSCT) based algorithm to detect the surface defects. Then, the detection results are sent to the mechanical separation system via a serial port. The defective capacitors are peeled off from the production line by motor. The proposed system can improve the detection efficiency. It thus can improve the product quality and reduce production costs. Experimental results have demonstrated that the system achieves superior performance over other state-of-the-art solutions. Moreover, with the system, large-scale vision data of capacitor surfaces can be collected and used to supervise capacitor manufacturing process.