DocumentCode :
1567039
Title :
Unmixing Component Analysis for Anomaly Detection in Hyperspectral Imagery
Author :
Gu Yanfeng ; Zhang Ye ; Liu Ying
Author_Institution :
Dept. of Inf. Eng., Harbin Inst. of Technol., China
fYear :
2006
Firstpage :
965
Lastpage :
968
Abstract :
Anomaly detection is one of the most important applications for hyperspectral images. In this paper, a new algorithm called unmixing component analysis (UCA) is proposed for anomaly detection in hyperspectral imagery. The proposed algorithm firstly performs spectral unmixing only with background endmembers on original hyperspectral images, and the unmixing error data are retained. Secondly, kernel principal component analysis (KPCA) is performed on the error data to concentrate and extract useful information about anomalous targets. After that, non-linear principal component that includes the most information about anomalous targets is selected based on non-Gaussianity measures. Finally, anomaly detection is conducted on the selected non-linear principal component using RX detector. Numerical experiments are performed on AVIRIS data with 126 bands. The experimental results show the proposed algorithm greatly modifies performance of the conventional RX algorithm and has good detection performance with low false alarms.
Keywords :
feature extraction; image processing; principal component analysis; AVIRIS data; KPCA; RX detector; UCA; anomaly detection; hyperspectral image; information extraction; kernel principal component analysis; unmixing component analysis; Algorithm design and analysis; Clutter; Detection algorithms; Detectors; Feature extraction; Hyperspectral imaging; Image analysis; Information analysis; Kernel; Object detection; anomaly detection; hyperspectral images; kernel principal component analysis; spectral unmixing;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Image Processing, 2006 IEEE International Conference on
Conference_Location :
Atlanta, GA
ISSN :
1522-4880
Print_ISBN :
1-4244-0480-0
Type :
conf
DOI :
10.1109/ICIP.2006.312648
Filename :
4106692
Link To Document :
بازگشت