多光谱图像的有效压缩已经成为遥感领域亟待解决的难题。针对星载多光谱成像仪获取的多光谱图像，提出了一种基于分块KLT（Karhunen-Loève transform, 卡胡南-洛维变换）的低复杂度有损压缩算法。该算法首先对每个波段分别进行空间二维小波变换，以去除多光谱图像的空间相关性；然后将每个波段分成互不重叠且大小相等的图像块，每次仅对相邻两个波段的对应图像块进行谱间KLT 变换，以去除谱间相关性；最后对变换后的所有波段进行联合EBCOT（Embedded Block Codingwith Optimized Truncation，最优截断的嵌入式块编码）压缩。实验结果表明，该算法的压缩性能优于基于整体KLT 的多光谱图像压缩算法，并且具有较低的编码复杂度。

Efficient compression of multispectral images has been a persistent problem in the field of remote sensing. As for the multispectral images captured by satellite, a block-based KLT lossy compression with low complexity is proposed. First, a two-dimensional discrete wavelet transform is performed on each band of the multispectral images to remove spatial correlation. Subsequently, each band is partitioned into non-overlapping blocks of the same size; blocks that are co-located on adjacent two bands are subjected to an adaptive Karhunen-Loève transform to remove their spectral correlation. Finally, the optimal truncation technique of post-compression and rate-distortion optimization is employed for rate allocation to multiple bands, followed by embedded block coding with optimized truncation to generate the final bit-stream. Experimental results show that the proposed algorithm not only outperforms the algorithm based on global KLT, but also has low encoder complexity.