高光谱图像异常检测算法研究进展  

Research progress of anomaly detection algorithms in hyperspectral images

作　　者：叶怡[1] 杨桄[1] 童涛[1] 

机构地区：[1]空军航空大学航空航天情报系,吉林长春130022

出　　处：《激光与红外》2013年第7期719-725,共7页Laser & Infrared

摘　　要：异常检测算法不需要利用光谱的先验知识,而能直接检测出与周围景物光谱存在明显差异的光谱信号所在位置,是高光谱遥感应用领域一个重要研究方向,在民用和军事上都有重要的理论价值和应用前景。深入分析了目前主要的异常检测算法,并对各算法的优、缺点分别进行了评述。最后,指出了此领域今后的研究方向。Anomaly detection algorithms do not require priori knowledge of spectrum, and can directly detect the loca- tion where the signal spectrum is distinct from that of the background. Anomaly detection has been an important re- search direction in the field of hyperspectral remote sensing applications. There are important theoretical values and prospects both in civilian and military applications. The current main anomaly detection algorithms are deeply ana- lyzed, and the advantages and limitations of each algorithm are also discussed. Finally, the direction of future research in this field is pointed out.

关 键 词：高光谱图像 异常检测 全局 局部 算法 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]