光学遥感压缩成像技术

Compressive Imaging Techniques in Optical Remote Sensing

基于Shannon采样定理的传统信息获取系统在高空间、时间和谱分辨率及系统其它性能上存在难以突破的瓶颈,压缩采样理论为提升航天遥感信息获取能力提供了新的思路。基于压缩采样理论的成像技术(压缩成像)将采样、压缩和数据处理3个过程完美的结合在一起,避免了传统遥感成像系统"先采样再压缩"方式带来的传感器和计算资源浪费,是未来光学遥感极具潜力的成像方式。文章在简要介绍压缩采样基本理论的基础上,总结和分析了国际上目前提出的光学压缩成像系统原型,设计开展了3组压缩成像物理实验,特别结合航天遥感需求设计了推扫式压缩成像方案,实验结果验证了压缩采样的基本原理,并为未来光学遥感压缩成像系统的设计提供了借鉴。

Compressive sampling provides a new way for increasing the capability of information acqui-sition. Compressive sampling asserts that it is possible to accurately reconstruct signals from sub-Nyquist sam-pling, provided we make some additional assumptions（sparse or compressible） about the signal in question. The compressive imaging technology, which is based on the compressive sampling theory, integrates the proc-esses of sampling, compression and processing perfectly, avoiding resource waste caused by a traditional“sam-ple-then-compress”mode, and is a potential imaging technique for optical remote sensing. This paper first re-views the basic theory of compressive sampling. Then, several optical compressive imaging systems are intro-duced. Finally three physical experiments are designed to validate the principle of compressive imaging and the experiment results can be used as reference for the future optical remote compressive imaging system.