星上高光谱图像无损压缩方法  

Lossless Compression Method for On-board Hyperspectral Image

作　　者：舒锐[1] 武少林 龚伟 陈阳[2] 张天序[2] SHU Rui;WU Shaolin;GONG Wei;CHEN Yang;ZHANG Tianxu(Shanghai Institute of Satellite Engineering,Shanghai 200240,China;School of Artificial Intelligence and Automation,Huazhong University of Science and Technology,Wuhan 430074,Hubei,China;School of Electrical and Information Engineering,Wuhan Institute of Technology,Wuhan 430062,Hubei,China)

机构地区：[1]上海卫星工程研究所,上海200240 [2]华中科技大学人工智能与自动化学院,湖北武汉430074 [3]武汉工程大学电气信息学院,湖北武汉430062

出　　处：《上海航天（中英文）》2021年第4期152-160,共9页Aerospace Shanghai（Chinese&English）

基　　金：上海航天科技创新基金(SAST2018-030)。

摘　　要：针对星上高光谱图像压缩问题,本文从消除高光谱图像的谱间与谱内空间冗余信息入手,提出了基于KLT与自由无损图像格式(FLIF)算法的高光谱图像无损压缩算法。KLT用于去除图像的谱间冗余信息,FLIF算法的MANIAC决策树能够在压缩过程中动态更新上下文环境内容,实时调整上下文环境的数量,多上下文环境建模的设计大幅提升了压缩性能。仿真实验表明:在高分五号卫星上的高光谱数据集上,本方案取得了不错的压缩比特率结果。此外,针对KLT变换后图像的特点,本文提出了多压缩器灵活切换的加速方法,在可见光波段让压缩耗时缩短为未优化前的1/5。A lossless compression algorithm based on the Karhunen-Loève transform(KLT)and the free lossless image format(FLIF)algorithm is proposed to eliminate the redundant information between and within the spectra of hyperspectral images,so that the problem of on-board hyperspectral image compression can be solved. The KLT is used to remove the redundant information between the spectra of the images. The MANIAC decision tree of the FLIF algorithm can dynamically update the content of the context environments and adjust the number of the context environments in real time during the compression process. The design of multi-context environment modeling greatly improves the compression performance. Simulation experiments show that the proposed scheme achieves good bit rate compression results on the hyperspectral data set of Gaofen-5 satellite. In addition,in view of the characteristics of images after KLT,an acceleration method for flexible switching of multiple compressors is put forward,which can shorten the time consumption of compression to one-fifth of that before optimization in the visible light band.

关 键 词：高光谱图像 无损压缩 可逆整数KLT 自由无损图像格式(FLIF) 

分 类 号：TN911.73[电子电信—通信与信息系统] TP391.9[电子电信—信息与通信工程]