摘要
本文根据高光谱图像具有空间和谱间相关性的特点，提出一种基于三维整型DCT变换的无损压缩方法。首先采用三维整型DCT变换消除高光谱图像空间和谱间的相关性；然后，对变换系数进行类似小波的树状系数重组，并按子带顺序进行一阶自适应算术编码。实验结果表明，本文提出的方法与JPEG2000中无损压缩算法相比，平均比特率降低0．1～0．4bpp；与JPEG-LS相比，平均比特率降低0．01～0．2bpp。
This paper presents a three-dimension integer reversible DCT scheme to compress the data losslessly according to the characteristic of hyperspectral data. First, the data are decorrelated spatially and spectrally using three-dimension integer DCT method, then DCT-based coefficients-reordering and Zigzag scanning by subbands are applied to improve the coding efficiency before the context-based adaptive arithmetic coding （AAC）. Experiments show that the performance of the present method is evident： 0.1-0.4bpp is saved with respect to JPEG2000, and 0.01-0.2bpp is saved with respect to JPEG-LS on average.
出处
《电路与系统学报》
CSCD
北大核心
2006年第2期30-32,29,共4页
Journal of Circuits and Systems
基金
北京市基金/市教委重点资助项目(KZ200310005004)
国家基金资助项目(60227101)
