一种综合的有噪图象边缘提取方法

A Hybrid Method for Edge Detection in Noisy Image

有噪图象的边缘提取是一个难点。经典的边缘检测方法由于引人了各种形式的微分运算，从而必然引起对噪声的极度敏感。为此，本文提出了采用分形方法并综合传统方法以解决低信噪比图象的边缘检测问题，并给出了自然场景中的人造物体边缘检测的实例。

Edge detection in a noisy image is considered as a difficult problem. The traditionaledge detection methods based on differential operation are very sensitive to the noise. For this reason,a new method based on fractal geometry and the traditional edge detection techniques is proposed for solving the problem of edge detection in low signal noise ratio,and a practiCal applicationof edge detection to a man-made object in natural scenes is made.