摘要
针对KRX方法对高光谱图像进行异常目标检测时存在检测效率低和虚警率高的问题,在充分分析高光谱图像数据特征基础上,本文提出一种最优波段子空间方法的高光谱图像异常目标检测算法。该算法首先利用双边滤波方法对高光谱图像进行全局滤波,充分利用双边滤波的优点,使得高光谱图像背景信息得到抑制;然后采用经典的自动子空间方法对高光谱图像进行波段子集划分;再利用联合偏度-峰度指标,在每个波段子集内选出最优波段;最后利用这些最优波段构成新的波段最优子空间,在此基础上,在最优波段子空间中利用Kernel RX算法进行异常目标检测,从而得到异常检测结果。本文利用真实的高光谱图像进行仿真验证,获得异常目标、检测的虚警数和ROC等检测结果。结果表明,该算法具有鲁棒性强、虚警率低和检测精度高等优点。
In order to overcome the low detection accuracy and high false alarm rate of anomaly target detection in hyperspectral image based on Kernel RX anomaly detector,the new hyperspectral image anomaly target detection algorithm for optimal bands subspace is proposed on the basis of fully analyzing the characteristics of hyperspectral image data.First,the bilateral filtering method is used to filter the hyperspectral image,which makes the background information of hyperspectral images be suppressed by full using of the advantages of bilateral filtering.Then,the band subset of hyperspectral image is divided by adopting the classical ASD method,and then the optimal bands are selected by using the combined skewness-kurtosis index in each band subset.Finally,the Kernel RX algorithm is used to detect anomaly targets by using these optimal bands.The real AVIRIS hyperspectral imagery data sets are used in the experiments,and the detection results are obtained by anomaly targets,false alarm number and ROC curve.The results show that the proposed algorithm has strong robustness and lower false alarm probability.
作者
成宝芝
张丽丽
CHENG Bao-zhi;ZHANG Li-li(College of Mechanical and Electrical Engineering,Daqing Normal University,Daqing 163712,China)
出处
《液晶与显示》
CAS
CSCD
北大核心
2019年第9期897-904,共8页
Chinese Journal of Liquid Crystals and Displays
基金
黑龙江省自然科学基金(No.LH2019F040)~~
关键词
高光谱遥感图像
异常目标检测
双边滤波
波段子空间
hyperspectral remote sensing imagery
anomaly target detection
bilateral filtering
bands subspace
