学术
摘 要:
针对高光谱图像中小目标检测问题，提出了一种基于约束能量最小化（Constrained Energy Minimization，CEM）的目标检测算法。该算法首先对原始图像进行背景信息抑制从而抑制背景地物、突出低概率的小目标，用迭代误差分析的自动端元提取算法找出目标的端元光谱，然后把目标端元光谱代入CEM滤波器得到该目标的检测结果图。用高光谱数据进行了实验研究，并与CEM滤波器进行了比较。结果表明，其检测性能与直接采用CEM方法的检测性能相当，但是相对于CEM方法，该算法不需要目标的先验光谱信息，更具有实用性。[著者文摘]
文章出处:
《光电工程》-2007年34卷7期 -18-21页
栏目信息:
分 类 号:
文献标识码:
A
文章编号:
1003-501X（2007）07-0018-04
Target detection algorithm in hyperspectral image based on CEM
XUN Li-na, FANG Yong-hua, LI Xin （ Anhui Institute of Optics and Fine Mechanics, the Chinese Academy of Sciences, Hefei 230031, China ）
Abstract:
A new target detection algorithm in hyperspectral image based on Constrained Energy Minimization （CEM） is introduced. The purpose of our target detection algorithm is to search for and locate the targets which are relatively small with low probabilities in an image scene. Firstly, the hyperspectral image was projected onto a subspace, which was perpendicular to the space spanned by transformation matrix of principal components. In this subspace, background information was effectively suppressed and small targets became visible. Therefore, end member spectra of targets could be extracted using iterative error analysis method. Then, using extracted spectra as the desired signature, CEM was implemented. Experimental results show that the algorithm can effectively and reliably detect the small target from hyperspectral image. The performance of the proposed algorithm is close to the performance of CEM, but the new algorithm is more applicable and does not need the prior information of the desired signal source.[著者文摘]
Key words:
remote sensing; target detection; CEM; hyperspectral image
更多>>免费文章
cqvip.com