采用KL-变换的三维谱象数据压缩方法

A data compression method for 3 dimensional spectral image data using KL transform

提出了一种基于ＫＬ－变换的三维谱象数据压缩方法。该方法将原始谱象分成若干子块，对每一子块使用去相关性效果最佳的ＫＬ－变换，去除谱象数据空间相关性。实验结果表明，经ＫＬ－变换后，若不编码，可获得３～５倍的准无损压缩。经ＪＰＥＧ编码后，压缩比可达１００倍以上，且峰值信噪比大于３４．０ｄＢ．

A data compression method for 3 dimensional spectral image data based on KL transforms is presented.In the method the original 3 dimensional spectral image data from imaing spectrometers are divided into sub blocks.Each sub block is transformed by KL transforms to remove spatial correlation.Experimental results show that near loss less compression ratio of 3∶1￣5∶1 can be obtained without coding,and loss compression ratio of 100∶1 can also be attained with JPEG codeing while peak signal to noise ratio(PSNR) is over 34 dB.