高光谱海量数据的有效压缩成为遥感技术发展中需要迫切解决的问题。该文提出了一种基于聚类的高光谱图像无损压缩算法。针对高光谱图像不同频谱波段间相关性不同的特点,根据相邻波段相关性大小进行波段分组。由于高光谱图像波段数量较多,采用自适应波段选择算法对高光谱图像进行降维,以获取信息量较大的部分波段,利用 k 均值算法对降维后的波段谱矢量进行聚类。采用多波段预测的方案对各组中的波段进行预测,对于各个分类中的每个像素,分别选取与其空间相邻的已编码的部分同类点进行训练,从而获得当前像素的谱间最优预测系数。对 AVIRIS 型高光谱图像的实验结果表明,该算法可显著降低压缩后的平均比特率。
This paper reports that an analytic method is used to calculate the load responses of the two-wire transmission line excited by a plane-wave directly in the time domain. By the frequency-domain Baum Liu-Tesehe （BLT） equation, the time-domain analytic solutions are obtained and expressed in an infinite geometric series. Moreover, it is shown that there exist only finite nonzero terms in the infinite geometric series if the time variate is at a finite interval. In other word, the time-domain analytic solutions are expanded in a finite geometric series indeed if the time variate is at a finite interval. The computed results are subsequently compared with transient responses obtained by using the frequency-domain BLT equation via a fast Fourier transform, and the agreement is excellent.
