一种高分辨率遥感影像的数学形态学滤波算法

A Mathematical Morphology Filtering Algorithm for High-Resolution Remote Sensing Image

应用数学形态学理论,结合高空间分辨率遥感影像的特点,选择包含全部方向的结构元素,设计一种基于全方位结构元素的加权多级滤波算法,用于高空间分辨率遥感影像的处理。算法中使用峰值信噪比方法自适应地确定各个方向结构元素的权重。实验结果表明:与结构元素尺度形状单一的形态学滤波算法相比,全方位结构元素的加权多级滤波算法在噪声抑制和细节保持上均有较好的性能,峰值信噪比有了更大的提高,而且更适合于对高空间分辨率遥感影像的噪声去除。

Based on mathematical morphology （MM） method and the characteristics of remote sensing image with high spatial resolution, a MM filtering algorithm using Omni-directional structure elements was developed to process remote sensing image. In this algorithm, the peak signal-to-noise ratio （PSNR） technique was placed to acquire the self-adaptive direction weightings. The results suggest that comparing to the MM filtering algorithm with simplex scale figure structure elements, the purposed algorithm can keep better properties in noise suppression and detail preserving, and the PSNR increases. This program is more effective in noise filtering and detail preserving, and is suitable for high-spatial-resolution remote sensing image.