Research On Efficient Compression Technology Of Satellite Multispectral Image

With spatial resolution and spectral resolution growth, huge data quantity puts forward higher requirements to performance, processing speed and reliability of multispectral image compression technology. It is urgent to propose higher performance compression technology through the research of new remote sensing image compression technology and making full use of multispectral image correlation.In this paper, the existing multispectral image compression technology is researched. On the basis of research on the existing international standard of image compression, some useful attempts are carried on. This paper firstly analyzes the characteristics of multispectral images, including spatial correlation, spectral correlation and information entropy. The Analysis of the characteristics of spectral image provides an important theoretical basis to propose the image compression algorithm adapted to the multispectral image. Then, this paper introduces two kinds of multispectral image preprocessing techniques:band reordering and image registration. Two algorithms laid a foundation for the next prediction. Finally, based on preprocessing technology and prediction of the multispectral image lossless compression scheme and based on preprocessing technology and edge detection of the multispectral image lossless compression scheme are proposed.Compared with several multispectral image lossless compression methods commonly used presently, the proposed programs have better compression effect on image registration and non-registration image. Considering the space-borne applications, the performance of complexity and anti-error of the two proposed algorithms is evaluated.