多光谱遥感图像压缩系统中的ROI编码方法

提出了一种基于EBCOT(率失真优化截取内嵌码块编码)算法的矩形ROI(感兴趣区域)编码的干涉多光谱卫星遥感图像压缩方法。该方法不需要对小波域的系数进行提升，而是在码流组织时通过对多光谱区域的误差跟踪提高恢复图像的质量。从而克服了传统方法因为增强系数与图像复杂度不匹配带来的ROI与BG的PSNR质量不协调的问题，该算法的解码器不需要知道该图像是否存在ROI，不需要反提升过程，完全正常解码即可，而且该方法保留了EBCOT的优良特性。实验表明，这种编码方式在干涉多光谱卫星图像压缩系统中可获得理想的效果。

A method for ROI coding in multispectral remote sensing image coding system

A new method of rectangle region of interest (ROI) coding for multispectral remote sensing images is proposed.The novelty is that the ROI coding does not lift the coefficients in wavelet domain but is implemented by code stream organization.It overcomes the disadvantage of the PSNR unbalance between ROI and BG (background) when a given lift coefficient is not suitable for the image.The method in the paper is based on Embedded Block Coding with Optimized Truncation (EBCOT),which is done by adaptive distortion tracing.The decoder does not need to reversely lift the coefficients and works normally.More importantly,this method remains all the features of EBCOT such as resolution progressive,PSNR progressive and good robust for error bits spread.Experiments indicate that this method can get a good effect in the multispectral image coding system.