基于混合滤波的遥感图像去噪

Noise Removal of Remote Sensing Image Based on Mixed Filter

遥感图像在生成、传输和转化过程中往往受到各种噪声的干扰,影响了图像本身的清晰度,降低了视觉效果,不利于真实信息的提取。研究目的是对高分辨率遥感图像进行预处理,以提高信息识别精度。研究方法是噪声检测结合混合滤波的方法,该方法基于均值滤波和中值滤波的互补性,首先对图像每一个像素点进行噪声检测,把受高斯噪声污染和受椒盐噪声污染的像素点区分开,然后分别对受高斯噪声污染和受椒盐噪声污染的像素点进行处理。与其他滤波算法进行比较的结果表明,噪声检测结合混合滤波方法能够有效去除椒盐噪声和高斯噪声,保留图像的纹理信息,提高图像的清晰度。

The RS image is often disturbed by lots of different noises during the procession of generation, transmission and transformation, which blurs the true information of the image and affects the visibility. The purpose of this study is to improve the precision of SPOT image, and the method used is a mixed filter combined with noise detection, which is based on the complementary of the average filter and the median filter. The method first classified the pixels into two classes： one is the pixels which corrupted by Gauss noise and the other by impulse noise, and than average filter is used for the pixels corrupted by Gauss noise while median filter is used for the pixels corrupted by impulse noise. Compared with other filters, the simulation result verifies that the method has a well performance on RS image, and the image detail is kept well simultaneously.