基于独立分量分析的高光谱图像压缩

Compression of Hyperspectral Image Based on Independent Component Analysis

在对原始数据进行虚拟维数估计的基础上,提出了一种基于最大距离端元提取+独立分量分析(Independent Component Analysis,ICA)的高光谱图像有损压缩方案.该方案首先应用最大距离端元抽取法提取高光谱数据各端元矢量,然后用快速独立分量分析生成独立分量,最后使用2维分层树集合分裂(Set Partitioning In Hierarchical Trees,SPIHT)算法对各独立分量图进行编码.计算机仿真结果证明,该算法在取得高压缩率的同时,能很好地保持数据的谱特征,是一种高效的三维数据压缩方法.

Based on the estimation of virtual dimensionality,a lossy compression method using independent component analysis （ICA） with given initial projection vectors is proposed. Endmember vectors are extracted using distance maximization method. Fast ICA is used and the ICA images are coded using 2-D set partitioning in hierarchical trees（SPIHT）. The experimental results show that the algorithm is an efficient method. High compression ratio is achieved and the spectral characteristics are preserved.