Lossless data compression for infrared hyperspectral sounders - An overview

Hyperspectral sounding data requires accuracy for useful retrieval of atmospheric temperature, moisture, trace gases, clouds, aerosols and surface properties. Therefore, compression of hyperspectral sounding data is better to be lossless or near lossless. Given the large volume of three-dimensional hyperspectral data that will be generated by the hyperspectral sounders such as AIRS, CrIS, IASI, GIFTS and HES instruments, the use of robust data compression techniques will be beneficial to data transfer and archive. This paper reviews wavelet-based lossless data compression schemes for the 3D hyperspectral data using 3D integer wavelet transforms followed by 3D EZW and 3D SPIHT zerotree coding schemes. We extend both zerotree coding schemes to take on any size of satellite data, each of whose dimensions need not be divisible by 2 ^N, where N is the layers of the wavelet decomposition being performed. The 2D wavelet-based JPEG-2000 compression scheme and some other prediction-based 2D lossless compression schemes such as CALIC and JPEG-LS are also investigated. Their compression ratios are presented.