Optimized-SSIM Based Quantization in Optical Remote Sensing Image Compression

High-rate compression usually causes serious distortion of texture and edges which play important roles in optical remote sensing image application. In order to reduce obvious structural distortion, in this paper, we analyze the correlation of SSIM (Structural Similarity) component functions with MOS (Mean Opinion Score) on an optical remote sensing compression distortion image database, conclude that SSIM should be substituted by its component function in optical remote sensing image compression assessment. After that, we utilize the component function to design a quantization approach, and apply it to an embedded wavelet image coder. Experiments show that our approach can preserve more structure and texture in image by high-rate compression.