基于网函数插值的高密度椒盐噪声去除方法  

Approach for Removing High Density Salt and Pepper Noise Based on Net Function Interpolation

作　　者：张晰[1] 张杰[1] 来全[2] 郎海涛[3] 张临杰[4] 

机构地区：[1]国家海洋局第一海洋研究所,山东青岛266061 [2]内蒙古师范大学,呼和浩特010022 [3]北京化工大学,北京100029 [4]中国海洋大学,山东青岛266100

出　　处：《计算机工程》2016年第8期220-226,232,共8页Computer Engineering

基　　金：国家自然科学基金资助项目"基于主被动遥感的渤海海冰厚度及其相关参数的反演研究"(41306193);海洋公益性基金资助项目(2013418025)

摘　　要：图像在获取和传输过程中容易产生椒盐噪声,严重影响图像质量。针对高密度椒盐噪声滤波处理难的问题,基于网函数插值,提出一种高密度椒盐噪声图像降噪方法。利用椒盐噪声的奇异点和不相关特性检测椒盐噪声,根据椒盐噪声所在位置,采用网函数插值方法对图像进行恢复和重构。实验结果表明,与经典的中值滤波和自适应中值滤波方法相比,该方法对于高密度(密度>70%)的椒盐噪声能取得更佳的滤波效果,更好地保持图像细节。Salt and Pepper（SP） noise is easily gererated in the process of image acquisition and transmission, which can seriously affect the image quality. For the difficult processing problem of high density SP noise filtering,an approach for removing high density SP noise of images based on net function interpolation is proposed. The singular point and irrelevant characteristics of SP noise are used to detect SP noise. According to the location of SP noise, net function interpolation method is used to restore and reconstruct the image. Experimental results show that the proposed algorithm has better filtering performance than the conventional median filtering method and adaptive median filtering method. The approach not only filters the high density （density 〉 70% ） SP noise but also maintains the image details.

关 键 词：网函数插值 椒盐噪声 图像降噪 高密度 图像处理 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]