大压缩比嵌入式遥感图像压缩

本文提出一种基于小波变换和矢量量化相结合的遥感图像压缩算法。该算法由于采用嵌入式的比特流构造方式，因而具有嵌入式编码（Embedded Coding)的特点，比特流按照重要性顺序发送，编码器可以在满足一定压缩比的情况下终止编码。在大压缩比的情况下，恢复图像具有良好的视觉效果；同时它具有比SPIHT算法优良的抗误码性能。

Remote sensing images via embedded coding in high compression rate

A new method which was used to compress remote sensing images is presented in this article.This algorithm was based on both wavelet-transform and vector quantization.After wavelet-transform,a vector was constructed with a spatial self-similarity between subbands.This approach remained the low frequency of wavelet coefficients in lossless mode,at the same time,reconstructed the detail information of wavelet coefficients efficiently.It made an embedded bit stream in order of importance so an encoder could terminate the encoding at the point that a target rate is met exactly.Although the peak-signal-to-noise-rate (PSNR) of this approach is close to that of SPIHT(Set partitioning in hierarchical trees coding,introduced by A.Said and WA.Pearlman,is a very effective and simple technique in images compression ),the reconstruction performance is better than that of SPIHT when transferring in a noise channel.The fact that error bits often result in extremely unnatural images suggests that SPIHT could not be used directly in remote sensing images compression in satellite communications.However,this algorithm will reconstruct the images with better performance in a noise channel.Based on all the virtue above,this algorithm is looking forward to being used in satellites.