基于谱间预测和码流预分配的高光谱图像压缩算法  

Hyperspectral image compression algorithm based on prediction between bands and rate pre-allocation

作　　者：孙蕾[1] 罗建书[1] 谷德峰[1] 

机构地区：[1]国防科技大学理学院,湖南长沙410073

出　　处：《光学精密工程》2008年第4期752-757,共6页Optics and Precision Engineering

基　　金：国家自然科学基金资助项目(No.60572135)[NNSF(No.60572135)]

摘　　要：在高光谱图像压缩中，由于波段预测后各波段残差图像分配的码流长度关系到该波段的清晰度，同时该波段作为预测波段对后面波段的预测编码会产生影响，因此设计一种合理的分配各波段码流长度的算法很有必要。本文提出了一种波段预测去除谱间冗余和码流预分配的高光谱图像压缩算法。用DPCM预测求出各波段的预测残差图像的标准差，然后根据标准差的大小对该波段进行SPIHT编码所需要分配的码流长度进行预测，最后基于均方差最小的线性预测器对图像各波段进行预测，根据事先分配的码流长度对各波段预测残差图像进行SPIHT编码。该算法在波段间采用最佳线性预测，有效地去除了高光谱图像的谱间相关性，同时，设计的分配码流长度的算法能够根据各波段信息量大小，以及与相邻波段的相关性来分配码流长度。实验表明，重建图像的平均PSNR比3D-SPIHT算法的高0．9～2．5dB，因此该算法对高光谱图像有损压缩非常适用。In hyperspectral image processing,the algorithm to allocate the rate of each band reasonably is important for image compression. In order to find a solution to transmission and storage of huge hyperspectral image data, a hyperspectral image compression algorithm based on the prediction between bands to remove spectral redundancy and the rate pre-allocation is proposed. The standard error of residual image of each band is computed by DPCM, and then the rate for SPIHT coding of each band is allocated by the value of the standard error. Finally, each band is predicted by a linear predictor based on minimal square error, and is coded by SPIHT algorithm according to the pre-allocated rate. The optimal linear prediction is used in this algorithm, so that the correlation between bands can be removed efficiently. Meanwhile, the rate pre-allocation algorithm can reasonably allocate rate according to both the information account of each band and the correlation between neighboring bands. Experiment results show that the average PSNR of this algorithm is higher 0.9-2.5 dB than that of 3D-SPIHT algorithm,which is very suitable for hyperspectral image lossy compression.

关 键 词：高光谱图像压缩 SPIHT算法 谱间预测 码流长度分配 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]