引入图像复原技术的遥感卫星图像压缩解码算法

Image Compression Decoding Algorithm Using Image Restoration for Remote Sensing Satellite

针对遥感卫星基于小波变换的光学图像有损压缩算法,提出了一种引入图像复原技术思路的算法,对遥感图像解码处理过程中获得的低频信息进行复原,可以获得更多的用于在频域范围内描述图像结构信息的高频分量增益,从而提升整个解码图像的质量。通过此算法与传统的小波压缩算法进行对比,结果表明：压缩恢复图像的峰值信噪比（PSNR）较传统算法有0.1～0.3dB的改善,压缩恢复图像表现出更为丰富的纹理特性,可降低对遥感卫星图像数据处理和天地链路数据传输的要求。

In this paper, lossy image compression based on wavelet theory is considered. An image restoration technology is introduced for wavelet compression. By applying image restoration to the low-frequency component obtained from entropy decoding in decompression process, a gained high-frequency component is retrieved, which is the expression of reconstructed image texture in frequency domain. As a benchmark, the algorithm presented is compared with the traditional wavelet compression algorithm. The results of the comparative experiments show that our algo- rithm performs better than traditional algorithms in PSNR, and the reconstructed image is more texture-richer than that by the traditional approach without restoration. Meanwhile, the require- ments of image data processing and transmission could be reduced.