【摘要】 本文介绍一种噪声图象的滤波方法 ,它不需要先验的图象模型 ,而且也不会损伤边界和细小的细节 ,在相加性和相乘性噪声的情况下 ,计算与边界方向有关的子集的局部平均值和方差 ,相应地沿着边界方向的噪声就被去掉 ,同时对边界有一定的增强效果更多还原

【Abstract】 The filtering method for noise image is introduced in the paper.This method is not needed the priori model of image,and it will not damage the boundary and its image details.The local mean and variance of the subset related to the boundary direction are calculated.In this way the noise along the boundary will be eliminated and some enhancement effect will appear in boundary.更多还原