基于小波变换的高分辨率遥感图像复原算法实现  

Implement of High Resolution Remote Sensing Image Restoration Algorithm Based on Wavelet Transform

作　　者：程伟[1] 杨世植[1] 文奴 崔生成[1] 

机构地区：[1]中国科学院安徽光学精密机械研究所中国科学院通用光学定标与表征技术重点实验室,安徽合肥230031

出　　处：《大气与环境光学学报》2015年第5期401-407,共7页Journal of Atmospheric and Environmental Optics

基　　金：国家科技支撑计划项目(2011BAB01B03)资助

摘　　要：为了促进遥感图像的后续研究,针对高分辨率遥感图像实现了基于小波变换的迭代收缩(IST)图像复原算法。考虑到算法在复原过程中对内存需求较大,实现过程中采用内存映射文件的方法,将高分辨率遥感图像映射到进程地址空间。针对分块复原图像时通常会伴有边缘跳变现象,影响拼接后的图像质量的问题,使用特殊分块策略对图像进行分块处理。复原算法在VC平台下实现,通过遥感图像复原实验,并对复原图像进行评价分析,复原性能和效率良好。In order to facilitate the subsequent research of remote sensing images, a high-resolution remote sensing image restoration algorithm based on wavelet transform and iterative shrinkage （IST） has been im- plemented. Considering that the restoration process in the algorithm needs large memory, memory-mapped file method is used to map the high-resolution remote sensing image into the process address space. Because of the edge jump phenomenon which affects the image quality after splicing image block, a special splicing strategy is used for block restoration. Restoration algorithm is implemented in the VC platform. Through remote sensing image restoration experiments, analysis and evaluation of image restoration, the restoration algorithm is proved to have good recovery performance and efficiency.

关 键 词：遥感图像复原 高分辨率 内存映射文件 分块策略 小波变换 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]