Enhanced quality LANDSAT image processing based on 4-level Sub-Band Replacement DWT

Satellite images are multi-spectral in nature that operate over wavelengths of wide range frequencies known as bands. Hence, they need to be processed cautiously. This paper presents a 4-Level DWT based Sub-Band Replacement (SR-DWT) Image Compression scheme, for LANDSAT satellite images. The algorithm is designed to compress various color band images captured by the satellite payload. By performing compression on these images and converting them to unique textured images, it exploits the sub-bands generated by the wavelet transform. Since the algorithm is a reversible process, it facilitates retrieval of the images back to colored format at the receiver. Due to the property of possessing spatial-frequency decomposition, the recovered images are of high quality with improved Peak Signal to Noise Ratio (PSNR) of 20dB and correspondingly very low MSE value. Using an astrophysical image database, the performance analysis of the 4-Level SR-DWT algorithm is illustrated.