Onboard multispectral data compression using JPEG-like algorithm: a case study

With ever increasing demand for high spatial and spectral resolutions, high number of bits of multispectral (MS) sensor imagery from space borne systems, but not compensated by an equivalent increase on onboard data transmission or memory limits, efficient data compression and/or streaming approaches gain importance. This paper discusses about the use of JPEG-Like algorithm, for which hardware and software were well proven from the Cartosat-1 spacecraft, to compress onboard high resolution multispectral imagery for future missions. It studies two possible ways of compressing the multispectral data: (1). Apply JPEG-like algorithm bandwise for all three bands, and decompress in ground processing. This would yield compression ratio (CR) of 1:3.31, (2). Combine IRS-Green and Red (since both are highly correlated bands) in quincunx sampling grid, compress the grid and IRS-NIR data by JPEG algorithm. This approach would have the advantage of a higher CR of 1:4.97. It was found that the JPEG like algorithm used in Cartosat-1 could be directly used for MS data onboard as it would still preserve the spatial and spectral contents of the multispectral information after decompression in ground processing. Further research work is required to improve the image quality in the latter case despite the fact that it offers a better CR.