In order to realize the automatic detection of the defects in the expansion valve mounting hole in industrial production,a defect defection method for expansion valve mounting hole was proposed based on machine vision.SURF feature matching was used to locate the region of interesting (ROI),and dual threshold hysteresis segmentation method was used to eliminate the noise;the contour edges were calculated by using the Canny operator,and the contour information was extracted by using the neighborhood growth method.In order to overcome the interference of defect information and accurately locate the center of valve hole,a method of circle center location based on geometric moment and least square fitting was proposed.According to the distance from the point on the contour to the center of the circle,the defect contour was segmented and marked on the input image,so the detection of defects was realized.The results show that by using the proposed algorithm,the defects of the expansion valve mounting hole can be accurately detected and the requirements of automatic detection in industrial production can be met.