基于局部多结构元素的遥感图像消噪算法

Denoising algorithm of remote sensing image based on local multiple morphological structure elements

提出了一种基于局部多结构元素形态学的遥感图像消噪算法。首先确定灰度突变像素点 ,利用定义的八种多结构元素对灰度突变像素点进行形态学的腐蚀操作 ,以实现对像素点的分类。若判断为噪声点则滤除该点 ,同时在该点处进行中值滤波以填补像素值 ;若判断为非噪声点 ,则保留该点像素值。仿真结果表明 ,该算法能够很好地滤除遥感图像中的孤立噪声点 ,保留遥感图像的细节 ,同时对参数的调整不敏感 。

A denoising algorithm of remote sensing image is presented in this paper. Those pixels with the gray value suddenly changing are determined. Then, total eight kinds of morphological structure elements are defined to implement the eroding operation at these pixels and complete classification of the pixels. Simulation results show that this algorithm can well remove the pulse noise in remote sensing image, meanwhile the image main details are correctly reserved. Furthermore, this algorithm is insensitive to the parameters and is applicable to many noised remote sensing images.