基于 Karhunen-leove 变换的多光谱数据压缩统计特性分析

The Characteristic Properties Analysis for Multispectral Data Compression Based on Karhunen-Leove Transformation

提出了基于Ｋａｒｈｕｎｅｎ－ｌｅｏｖｅ变换（ＫＬＴ）的多光谱图像数据压缩方法，并对多光谱图像数据ＫＬＴ统计特征进行了分析。在对ＫＬＴ进行了理论推导后，如何将该方法应用于实际多光谱数据压缩系统进行了分析讨论。研究提出了一种最大限度地减少额外码流开销的方法。实验结果表明：该方法对１６光谱段的三维数据可获得３．２～５．３的压缩比，且几乎没有失真，明显优于用ＤＣＴ去除谱相关性的方法。

In this paper a compression method for multispectral image data basedon Karhunen-Leove Transformation (KLT) was presented, and the statistical properties of multispectral image using KLT was analyzed. After deducing the theoryof KLT, we discussed the schemes of applying this method to the practical compression system of multispectral image data. A maxium degree decreasing bit flow method is researched and presented. Experiment results shows that this methodcan obtain the compression ratio 3 2～5 3 for sixteen band multispectral, almostlossless. It is superior to the method of applying DCT to remove the spectral correlation .