Vision inspection of test button of electric leakage circuit breaker
XIE De mao, WU Zi ran, WU Gui chu, ZHAO Sheng
(College of Physics and Electronic Information Engineering, Wenzhou University, Wenzhou 325000, China)

Abstract: Aiming at automatic detection of leakage circuit breakers’ T button (test button) installation direction, an inspection system was designed. According to the environment and detection requirements, the appropriate camera、lens and light mode were determined, which obtain proper image that suitable for further processing. To solve the belt seriously jitter problem, the RGB color image space of the leakage circuit breaker and component labeling algorithm was studied, and a T button location algorithm about labeling color difference image was introduced. Automatic extraction of T button image is realized. For the sake of common binarization algorithm cannot achieve better segmentation of the T button, a binarization method based on normalization was proposed. The algorithm, which utilizes the maximum and minimum pixel values of the image, is adaptive. Finally, character template matching was utilized to detect the installation direction of the T button. The results indicate that the system achieves high reliability and efficiency, the detection accuracy rate reaches 99% and the false detection rate is lower than 1%, which meets the practical production requirement.

Key words: machine vision; leakage circuit breaker; test button; connected region labeling