Given that the vision technology can accurately identify and position target bearing in its application of bearing quality inspection and size measurement,a bearing identification and location algorithm based on machine vision was proposed.By preprocessing the collected bearing image,the target image was segmented and the outer contour edge features of the image were extracted.The template image of the bearing was set,and by combining with Hu moment invariant features of the image,the bearing was recognized and matched.The circle fitting of the image edge points was set by the least square method,and the iterative method was adopted for correction.By calculating the position coordinates of the center of the circle,the positioning of the bearing was realized.The experimental results show that the recognition matching degree of vehicle bearing is between 0 and 0.03,and the positioning error is within 0.5 pixel,which meets the requirements of the system for the accuracy of bearing recognition and positioning.
关键词
机器视觉/图像处理/Hu不变矩/最小二乘法
Key words
machine vision/image processing/Hu moment invariant/least square method
维普
万方数据