DCT and wavelet based image compression in satellite images

Synthetic aperture radar (SAR) and SPOT images are becoming increasingly important and abundant in a variety of remote sensing and tactical applications. Thus, there is a strong interest in developing data encoding and decoding algorithms that can obtain higher compression ratios while keeping image quality to an acceptable level. This study aims to compare most of the well-known compression techniques namely discrete cosine transform and discrete wavelet transform. It investigates RADARSAT and SPOT images of different regions of different characteristics. RADARSAT-1 fine and standard beam mode images of the different regions in Istanbul are used to test the performance of the compression algorithms. The regions, which have been investigated, were sea areas, forest areas, built environment-residential and industrial areas which define different patterns of urban land use. The studies showed that homogeneous areas like forest and sea gave better compression results compared to heterogeneous areas like industrial and environmental. The second purpose of this study is to compare the two compression algorithms. The discrete wavelet based algorithm gave much better results compared to the discrete cosine transform based algorithm. The results changed according to the quantization process and the transformcoding algorithm.