32- Band Hyper-spectral Image Compression using Embedded Zero Tree Wavelet

Hyper-spectral Image is a three dimensional array containing spatial (image) information on x & y axis and spectral information on z axis. It is high resolution imagery. Hyper-spectral image is a dynamic & powerful method to visualize data in terms of spatial & spectral features. It consumes a huge amount of storage. So due to large memory requirements, Hyperspectral images demand efficient compression. The Hyper-spectral image which we have taken in this paper is the Moffett field image which is taken by Airborne Visible Infra Red Imaging Spectrometer (AVIRIS) satellite. It consumes about 16 GB of space in a day. In this paper we have presented an algorithm which adapts EZW to 3D tree for quantization, Wavelet Transform for decorrelation and Entropy coding is done by Adaptive Arithmetic Coding. The results obtained show the efficient compression and are comparable to JPEG 2000 standard.