Abstract
Although several hyperspectral anomaly detection algorithms have proven useful when illumination conditions provide for enough light, many of these same detection algorithms fail to perform well when shadows are also present. To date, no general approach to the problem has been demonstrated. In this paper, a novel hyperspectral anomaly detection algorithm that adapts the dimensionality of the spectral detection subspace to multiple illumination levels is described. The novel detection algorithm is applied to reflectance domain hyperspectral data that represents a variety of illumination conditions: well illuminated and poorly illuminated (i.e., shadowed). Detection results obtained for objects located in deep shadows and light–shadow transition areas suggest superiority of the novel algorithm over standard subspace RX detection.
© 2010 Optical Society of America
Full Article | PDF ArticleMore Like This
Asif Mehmood and Nasser M. Nasrabadi
Appl. Opt. 49(24) 4621-4632 (2010)
A. V. Kanaev, B. J. Daniel, J. G. Neumann, A. M. Kim, and K. R. Lee
Opt. Express 19(21) 20916-20929 (2011)
Michael T. Eismann, Joseph Meola, Alan D. Stocker, Scott G. Beaven, and Alan P. Schaum
Appl. Opt. 47(28) F27-F45 (2008)