基于波段选择的RX改进算法  

Improved RX Algorithm Based on Band Selection

作　　者：彭娜 PENG Na(Shanghai Aerospace Electronic Technology Institute,Shanghai 201109,China)

机构地区：[1]上海航天电子技术研究所,上海201109

出　　处：《上海航天（中英文）》2022年第2期72-75,共4页Aerospace Shanghai（Chinese&English）

基　　金：装备预研基金(9140A21010114HT90065)。

摘　　要：针对高光谱检测亚像元飞机的问题,提出基于大气吸收谱段的改进RX异常检测算法,实现亚像元检测能力和检测效率的提升。阐述了改进算法的基本原理,对比了10 km高度和地面大气吸收系数的曲线差异,建立了改进算法的检测模型,采用仿真的客用飞机可见-短波高光谱数据和海水背景高光谱数据,获得了信噪比10 dB的高光谱仿真图像。采用经典RX算法和基于4个大气吸收波段的改进RX算法,对150个谱段仿真图像进行检测,获得了2种算法亚像元的检测能力和受试者工作特征(ROC)曲线。结果显示,改进RX算法能够有效提升亚像元检测能力,缩短检测时间。In order to solve the problem of hyper-spectral detection of sub-pixel aircrafts,an improved RX anomaly detection algorithm based on atmospheric absorption spectra is proposed to improve the detection capability and detection time efficiency of the sub-pixels. The basic principle of the improved algorithm is expounded,and the difference curve of the atmospheric absorption coefficient between the 10 km altitude and the ground is research. The detection model of the improved algorithm is established. A hyper-spectral simulation image of a signal-to-noise ratio of 10 dB is obtained with the visible-short wave infrared hyper-spectral data of passenger aircrafts and the sea water hyperspectral data by simulation. The classical RX algorithm and the improved RX algorithm based on 4 atmospheric absorption bands are used to detect 150 spectrum simulation images,and the detection capability and receiver operating characteristic(ROC) curves of the sub-pixels are obtained with the classical and improved algorithms. The results show that improved RX algorithm can effectively improve the detection capability of the sub-pixels and shorten detection time.

关 键 词：高光谱 检测 异常检测 RX算法 大气吸收 

分 类 号：V19[航空宇航科学与技术—人机与环境工程]