CCSDS压缩算法对高光谱数据质量的影响研究  

Influence of the Hyperspectral Data Quality Based on CCSDS Compression Algorithm

作　　者：朱学伟[1,2] 李晓辉[1] 朱博[1] 

机构地区：[1]中国科学院光电研究院定量遥感应用技术重点实验室,北京100094 [2]中国科学院大学,北京100049

出　　处：《遥感信息》2013年第4期29-36,共8页Remote Sensing Information

基　　金：国家科技支撑计划(2011BAH23B0)

摘　　要：成像光谱仪能够探测获取目标的空间信息和光谱信息,逐渐在军事/民用遥感领域广泛应用。然而随着成像光谱仪的空间分辨率和光谱分辨率的提高,数据量也飞速提高。受数据下行链路带宽限制,星载高分辨率成像光谱仪所获取的海量数据必须进行有损压缩,而采用有损压缩又带来了一个关键问题:有损压缩所造成的数据失真究竟会对高光谱数据质量及后续遥感应用造成怎样的影响。本文基于CCSDS压缩算法的两种压缩方案,从统计性能、辐射性能、空间性能、光谱性能和应用性能5个方面,系统性分析了数据压缩对高光谱数据质量造成的影响。结果表明,利用高光谱数据的高谱间相关性,采用谱间去相关与CCSDS空间数据压缩相结合的方案,与直接采用CCSDS进行空间数据压缩的方案相比,具有更好的压缩性能,对高光谱数据质量造成的影响更小。Hyperspectral imagery is capable in a wide diversity remote sensing applications because it can provide direct i- dentification. The hyperspectral image consists of a series of spectral bands, so the size of image is so large that the data must be lossy compressed to alleviate the downlink communication. However,there arises a key problem, that is how data distortion in- fluences the quality of hyperspectral image,which is caused by lossy compression. Based on the statistical performance, radiant performance,spatial performance,spectral performance and applied performance of hyperspectral image, this paper studied the two CCSDS compression schemes. Experiments show that the performance, which in five aspects of hyperspectral image quality, of compression scheme which combined with spectral decorrelation and CCSDS spatial compression,is better than scheme that is CCSDS spatial compression.

关 键 词：高光谱 CCSDS 图像质量 谱间去相关 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]