基于对偶树复小波变换的遥感图像薄云去除  

Thin cloud removal of remote sensing image based on dual-tree complex wavelet transform

作　　者：胡根生[1] 黎晓伊 梁栋[1] 黄林生[1] 

机构地区：[1]安徽大学电子信息工程学院,安徽合肥230601

出　　处：《安徽大学学报（自然科学版）》2016年第1期50-58,共9页Journal of Anhui University(Natural Science Edition)

基　　金：国家自然科学基金资助项目(61172127);安徽省自然科学基金资助项目(1408085MF121)

摘　　要：在卫星传感器获得的遥感图像中,由于受气候影响,图像可能存在被云雾噪声覆盖的区域.提出一种基于对偶树复小波变换的薄云去除方法,利用该变换将含云遥感图像分解成高频和低频成分,通过对高频补偿和低频抑制处理,有效去除遥感图像中的薄云,恢复云覆盖区域的地物信息.实验结果表明,该方法优于常用的薄云去除方法,在去除薄云的同时能有效恢复云覆盖下的地物信息,此效果源于对偶树复小波变换具有的近似平移不变性和良好的方向选择性.Remote sensing images obtained by satellite sensors are often covered with noise of clouds which result from the impact of climate.A thin cloud removal algorithm based on dual-tree complex wavelet transform was proposed in this paper.Remote sensing image covered with clouds was decomposed into high and low frequency components by using this transformation firstly.Then the high frequency components were compensated and the low frequency components were suppressed.The thin cloud on remote sensing image was effectively removed and the ground information in the cloud covered areas was recovered.Experiment results showed that,due to the performances of approximate translation invariance and good directional selectivity of dual-tree complex wavelet transform,the algorithm proposed in this paper was superior to the commonly used algorithms for thin cloud removal,and it could effectively restore ground information in the cloud covered areas.

关 键 词：遥感图像 薄云去除 图像增强 对偶树复小波变换 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]