多光谱影像星上压缩方法探讨

在星地数据传输信道带宽有限的情况下,为提高星载影像数据的单轨下传量,科研人员已开展卫星影像数据星上压缩方法的研究工作。区别于全色影像的压缩,多光谱或高光谱影像的星上压缩不仅需要考虑诸如JPEG等基于空间相关性的压缩算法,还需要结合如PCA等基于谱间相关性的方法,从而在数据压缩过程中保证空间和光谱两方面的信息保真度。因此,本文将回顾目前基于空间变换和谱间变换,以及基于内容的遥感影像压缩方法。进一步针对常规性监测和灾害应急监测两种需求,在考虑星上数据预处理和智能处理的基础上,提出了一种多光谱或高光谱影像星上压缩的理论框架及其对应的地面数据处理流程。最后,本文指出了多光谱或高光谱影像星上压缩研究工作中需进一步探讨的若干技术问题。

Discussion on the Onboard Compression Methods of Multi-spectral Images

When the bandwidth of data transmission channel is limited,to increase the data volume of the satellite image acquired in the single pass,the research on onboard compression of image data has been undertaken.Different from the compression of panchromatic images,for onboard compression of multi/hyperspectral image,it is critical to consider not only the compression methods based on spatial transform such as JPEG,but also the methods based on spectral transform such as PCA so as to guarantee the information fidelity of the image in both spatial and spectral domains.Therefore,current image compression methods based on spatial and spectral transform and contents are reviewed.Furthermore,to meet the requests of routine monitoring and disaster relief,a theoretical framework of onboard compression of multi/hyperspectral images and a workflow of ground data processing are proposed.Finally,the technical issues related to development of onboard compression methods of multi/hyperspectral images are also addressed.