Multispectral image compression for future LANDSAT remote sensing systems

Multispectral image compression has been identified as a critical technology area for future advanced land remote sensing systems. The nature of multispectral imagery is such that, with correlated spectral bands, additional redundancy exists between registered pixels which can be exploited for compression gains. Because of the stringent requirements placed upon multispectral compression by various exploitation activities, designing a multispectral compression algorithm is not trivial. This paper will concentrate, not on compression algorithms, but upon some of the system design requirements, image properties and other issues that face compression algorithm designers