摘要： 针对传统形态滤波器滤除高浓度椒盐噪声不足的问题,提出一种基于形态开闭算子自适应的高浓度椒盐噪声去除方法。该方法分为噪声检测和形态开闭自适应滤波2 个阶段。在噪声检测阶段,得到噪声标记图像,并依据噪声标记图像生成自适应的结构元素。在形态滤波阶段,对可能的噪声点进行形态开闭滤波,而对非噪声点不做滤波处理直接输出。通过一个简单的噪声检测方法构造自适应结构元素,提升传统形态滤波器的滤波能力。 仿真结果表明,该方法能有效去除高浓度的椒盐噪声,较好地保留图像的细节信息,并且算法简单,运算时间较短。

关键词: 椒盐噪声, 数学形态学, 开算子, 闭算子, 滤波, 自适应结构元素

Abstract: According to shortcoming of conventional morphological filter in filtering high density salt and pepper noise of digital images,a method for removing high density salt and pepper noise based on morphological open and close operators is proposed. This method consists of the noise detection stage and adaptive morphological open-close filtering stage. A mark image is constructed in noise detection stage,and adaptive structuring elements are gained based on the mark image. The possible noise pixels are filtered by morphological open and close filtering,and the free-noise pixels are unchanged in morphological filtering stage. Simulation results show that the proposed method is effective to remove high density salt and pepper noise,and keeps the detail information well. The proposed method constructs adaptive structuring elements using noise detection,improves traditional morphology filtering ability,and the algorithm is simple,possesses the shorter processing time.

Key words: salt and pepper noise, mathematical morphology, open operator, close operator, filtering, adaptive structuring element

中图分类号: 

• TN911. 73