小波变换在遥感图像云雾处理中的应用

Thin Cloud and Mist Reduction from Remote Sensing Images Based on Wavelet Transformation

分析受云雾影响的遥感图像频率分布的特征,云雾处于相对低频、景物处于相对高频。根据小波变换多分辨率特点,图像经多层小波分解,得到的低层细节系数代表图像的相对高频部分,高层细节系数代表图像的相对低频部分。提出通过增大图像的高频细节系数,减小低频细节系数,保持最低层近似系数,达到去云雾目的。利用视觉评价、均值、标准差、熵、平均梯度等方法评价实验结果,表明算法的有效性。

By analyzing the frequency distribution characteristics of the remote sensing multispectral image influenced by thin cloud and mist in both the theory and the practical application aspects, we conclude that the underlying bed detail coefficient represents the relatively high-frequency of the image, the high level detail coefficient represents the relatively low frequency band of the image. This paper proposes a novel method which can effectively strengthen the high-frequency component of an image and weaken its low frequency component based on the characteristics of the multi-resolution of the wavelet transform, and achieve goal of thin cloud and mist reduction. The experimental results the proposed algorithm are evaluated by using visual perception, mean value, standard differences, entropy, and average gradient, and are found to be satisfactory.