一种结合空谱聚类的高光谱图像快速压缩算法

对高光谱图像进行快速压缩已经成为了高光谱遥感领域的研究热点。针对现有的高光谱图像数据量大和压缩所需运算量大的问题，提出了一种基于频段聚类+主成分分析(PCA)与空间分类相结合的高光谱图像快速压缩算法。首先利用最大相关度频段聚类算法(MCBC)将频段聚类，接着将每一类频段用PCA压缩，然后将压缩后的图像利用聚类信号子空间投影(CSSP)算法进行图像分类，最后在每一类内利用LBG(Linde Buzo Gray)算法通过矢量量化快速完成高光谱图像的编码。在不同的压缩比下进行实验，结果表明提出的高光谱图像压缩算法能在保证良好的图像恢复质量的前提下，大幅度降低运算复杂度，实现高光谱图像的快速压缩。

A fast hyperspectral image compression algorithm combined with group and region clustering

Fast compression of hyperspectral image has become a hot topic in the field of hyperspectral remote sensing.Owing to the large amount of hyperspectral image data and the large amount of computation required for hyperspectral image compression,this paper presents a fast hyperspectral image compression algorithm based on band clustering+principal component analysis(PCA) and image segmentation.Firstly,the algorithm clusters the bands using maximum correlation band clustering(MCBC)algorithm.Secondly,it compresses each band cluster using PCA.After compression,the clustering signal subspace projection(CSSP) algorithm divides the image into proper regions.Finally,it finishes the encoding of each image region by vector quantization using Linde Buzo Gray(LBG) algorithm.The simulation results under different compression ratios show that the proposed algorithm can achieve a significant reduction in computational complexity and rapid compression of hyperspectral images,while ensuring good quality image restoration.