Multispectral band image compression using wavelets

Multi-band satellite imagery used in remote sensing for environmental monitoring, mapping and land use planning is compressed using Discrete Wavelet Transform to reduce storage capacity and transmission bandwidth in an efficient and effective manner. This paper explains the use of 2D Discrete Wavelet Transform for compression of Landsat5 multispectral band satellite image and analysis of its performance using various parameters such as Compression Ratio (CR), Peak Signal to Noise Ratio (PSNR), Digital Number Minimum (DN min) and Digital Number Maximum (DN max). Various wavelets are applied over the multi-band image for compression and the performance is analyzed. Using those values, the Spectral Signature graph is plotted and a particular wavelet is selected based on the parameter values.