基于滤波反投影的二值图像边缘检测新方法

A new edge detection algorithm for a binary image based on filter back projection

利用基于滤波反投影(FBP)的CT图像重建算法中出现的"杯状效应",对重建得到的CT值中的最小负值设置阈值因子来调整图像的整体归一化结果,从而得到一种新的边缘检测方法;同时给出一种改进的距离准则,以便针对不同的图像选择最优的阈值因子使得边缘检测的效果达到最佳。

Using the so called ＂cup effect＂appearing in CT-image-reconstruction based on filter back projection, a new method of edge detection has been obtained by adjusting the result of image normalization with a given threshold factor to the minimum minus values of the reconstructed CT-values. Furthermore, through an im- proved distance rule, the optimum threshold value which gives the best edge detection in the different images, is presented.