对于一些图像来说，常用的边缘检测算法有时候无法设立合适的阈值将它们的梯度较小的模糊边缘检测出来。为了解决这个问题，有两种解决方法：将图像方差标准化，拉大模糊边缘的梯度值，或通过设置sigmoid函数，将像素所在区域的信息传递到梯度值中去，对其进行调整，就能找到合适阈值，有效地将模糊边缘提取出来。本程序把这两种算法实现并与与传统算法进行了比较。-For some images, the commonly used edge detection algorithm sometimes can not be an appropriate threshold for the gradient of their smaller fuzzy edge detected. To solve this problem, there are two solutions: the image variance standardization, widening the fuzzy edge of the gradient value, or by setting the sigmoid function pixel region to disseminate the message to the gradient values to adjust, on the can find a suitable threshold, effectively extracted fuzzy edge. This procedure of these two algorithms with the traditional algorithm.