摘要
本文利用图像的相关性原理,结合图像的边缘信息,提出了一种新的图像边缘保持的方向平滑算法。该算法主要通过控制平滑区域大小和均方差阈值来选择受噪声干扰最小和最大相关区域。仿真证明该方法能够较好地去除椒盐噪声和高斯噪声,并且能够很好地保持图像的边缘和细节信息,作为图像的一种预处理方法,具有较高的实用价值。
In this paper, a new image＇s edge preserving orientation smoothing algorithm is proposed according to image＇s correlation theorem and edge characteristic information. The algorithm mainly controls the size of the smoothing area and the mean variance threshold to select the maximum correlation area containing minimum noise interference. The simulations in this paper prove that it can perfectly restrain salt ＆ pepper noise and gaussian noise by using the algorithm, meanwhile, the image＇s edges and details are well preserved. As an image pre-processing method, it is of good practical value.
出处
《计算机科学》
CSCD
北大核心
2006年第10期210-212,共3页
Computer Science
基金
重庆市自然科学基金资助(No.2005BB2065)。
关键词
边缘保持
方向平滑
最大相关邻域
均方差阈值
Edge preserving, Orientation smoothing, Maximum correlation area, Mean variance threshold
