基于区域信息的模糊加权图像恢复方法  

Region-information Based Fuzzy Weighted Mean Approach for Image Restoration

作　　者：王保平[1] 李宗领 谢维信[3] 

机构地区：[1]西安电子科技大学电子工程学院,西安710071 [2]咸阳师院教务处,咸阳712000 [3]深圳大学校长办公室,深圳518060

出　　处：《计算机工程》2003年第19期11-12,108,共3页Computer Engineering

基　　金：国家自然科学基金资助项目(69972041)

摘　　要：传统滤波器的优点是简单快速和不需图像的先验知识,但其效果不佳。neuro-fuzzy滤波器的优点是滤波效果较好,但滤波时需要一段训练学习时间。针对这点,文章提出了一种新的模糊滤波器,该滤波器除了具有上述两类滤波器优点的同时,又避免了它们各自的不足。实验表明该方法滤波效果优于传统的滤波器和其它模糊滤波器。Conventional filter has the benefits that it is simple and quick, but its performance is not good. Neuro-fuzzy filters has the benefits that its performance is good, but it needs a lot of training and learning time. In the light of this, a new fuzzy filter is proposed in this paper. The filter has advantages of both kinds of filters besides avoiding shortcomings of them. Experimental result shows that new-filter gives superior performance compared with conventional filters and other fuzzy filtes.

关 键 词：模糊加权平均滤波器 脉冲噪声 区域信息 图像恢复 模糊参数 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]