一种适用于多个可转位刀片的周边刃边缘检测方法

A Peripheral Edge Detection Method for Multiple Indexable Inserts

针对可转位刀片周边刃边缘检测的问题,提出了一种适用于多个可转位刀片的基于局部模糊检测和双树复小波变换的边缘检测算法。该算法主要分为两步,第一步是先按事先规定的要求拍摄可转位刀片的图片,然后对拍好的图片进行局部模糊检测处理,其中的检测算法是以图片中像素的相关性系数和基于Laplacian算子的新型判别算子为评价指标判断其是否位于模糊区域(若位于模糊区域,则用白色显示它,反之,则用黑色显示)。模糊检测算法在剔除误检像素后最后会生成一副二值图,可转位刀片的周边刃边缘即为二值图中黑色区域和白色区域的交界线。第二步就是用双树复小波变换对二值图进行边缘检测,检测后的结果就是我们需要的结果。实验结果证明,该算法能准确快速地检测出多个可转位刀片的周边刃边缘。

As for the problem of peripheral edge detection for indexable inserts , this paper propose an algorithm of edge detection based on local-blur detection and dual-tree complex wavelet transform for multiple indexable in- serts. The algorithm have two steps： the first step is to take pictures of indexable inserts according to the require- ments specified in advance firstly,and then make local-blur detection in the picture which was been taken,here the algorithm of edge detection judge whether the pixel is in the blur area with the correlation coefficients of the pixel in the picture and the new operator based on laplaeian operator as evaluation indexs （if it is in the blur area, show it in white; otherwise, show it in black ）.After eliminating the mistake pixels, the last step of the algorithm of edge detection will create a binary image,the peripheral edge of indexable inserts are the demarcation lines be- tween white areas and black areas in the binary image.The second step is to detect the edges in binary image by using dual-tree complex wavelet transform,the result ater transform is that we need. Experimental results show that this method can detect the peripheral edge of multiple indexable inserts quickly and accurately.