We present a new approach for compressing interference multispectral remote sensing images. A new rectangle region-of-interest (ROI) coding algorithm is integrated with an embedded-block-coding optimized truncation (EBCOT) method to make a distortion tracking of the spectral region. Accordingly, a rate-distortion optimizer is designed to give an adaptive lifting of the rate-distortion slopes. This allows the peak signal-to-noise ratio (PSNR) of the reconstructed multispectral region to be guaranteed higher than 38 dB for different image conditions. In this coding scheme, the lifting procedure is carried out in a code stream organization stage instead of after the DWT, as suggested by JPEG2000. As a result, the coarse lifting step situation due to the nature of bitplane shift methods can be improved, which enables flexible adjustment of compression fidelity in the ROI and background (BG), and improves the quality of the BG and the whole image. The Experimental results are encouraging.
River Jiang Xiao,
"Interference multispectral image compression using a new JPEG2000 region-of-interest coding method," Optical Engineering 43(4), (1 April 2004). https://doi.org/10.1117/1.1666859