【摘要】 提出了一种基于机器视觉的四方垫片内圆缺陷检测方法。首先利用广义Hough变换定位垫片的中心点位置,然后根据设定阈值对原图进行二值化、标记连通区域并删除最大连通区域外的区域,再分段搜索垫片内圆边沿点,最后用最小二乘法拟合圆后检测各边沿点是否为缺陷位置。算法对任意摆放位置和角度的四方垫片都能成功检测,实时性、准确率都可以满足工业检测的需求。更多还原

【Abstract】 A method for detecting inner circle defects of square gasket based on machine vision is proposed. First of all, locate the center point position of the gasket by using generalized Hough transform. Then, binarize the original image according to the set threshold, label the continuous areas and remove all areas other than the biggest. After that, search the edge points of inner circle of the gasket section by section. At last, fit the circle using least square method and detect whether the edge points are defect position or not. The algorithm can successfully detect the square gasket at any position and angle, and both real-time and accuracy can meet the needs of industrial inspection.更多还原