关键词: 机器视觉, 缺陷检测, 磁瓦, 图像处理

Abstract: An automatic arc segments ceramic magnet surface defect detection system based on machine vision technology has been developed.In the system，many image processing technologies have been adopted.The image acquisition system is designed according to the features of functional surface，and the improved mean filtering technology is used to eliminate noises，and the threshold segment and binarize based on gray histogram is used，and defect outline is distilled by morphology processing and Roberts edge detecting，and product quality level is decided by distilling defect area characteristic and template matching pattern recognizing method.Product line running indicates that the system takes on characteristics such as system running status is stable，detecting precision is high，which overcomes the disadvantages of large labour and low precision by handwork detecting.

Key words: machine vision, disfigurement detecting, arc segments ceramic magnet, image processing