陶瓷基片检测中Robert边缘细化算法

Edge thinning based on robert operator in detection of ceramic substrates

陶瓷基片检测中为了对亚像素进行精确定位,需要对基片图像进行灰度边缘细化.为了得到保留信息的细化灰度边缘,对边缘细化算法进行了研究,提出了Robert边缘细化算法.先对基片图像进行均值滤波,然后用Robert算子进行两次边缘检测,最后将两次检测结果进行减法运算得到边缘细化的陶瓷基片图像.实验表明,该算法可以得到允许范围内的单像素连通的细边缘,保留边缘灰度信息,为下一步对图像进行亚像素细分,提高测量精度打下良好的基础.

In order to detect the accurate position in the sub-pixel level during the detection of ceramic substrates,the gray-scale edge thinning of substrates is needed.In order to get the thin gray edge that remains the information,the edge thinning algorithm is studied,and the edge thinning algorithm based on Robert operator is proposed.First,mean filter is executed on substrate image.Then,twice edge detection is executed on the image.Finally the twice detection results were subtracted,getting the edge thinning ceramic substrate image.The experiment results show that the algorithm can supply the connected single-pixel thin edge within the scope,which lays a good foundation for improving the measurement accuracy during the sub-pixel positioning.