Linear and kernel methods for multivariate change detection

Morton John Canty, Allan Aasbjerg Nielsen

AbstractThe iteratively re-weighted multivariate alteration detection (IR-MAD) algorithm may be used both for unsupervised change detection in multi- and hyper-spectral remote sensing imagery as well as for automatic radiometric normalization of multitemporal image sequences. Principal components analysis (PCA) as well as maximum autocorrelation factor (MAF) and minimum noise fraction (MNF) analyses of IR-MAD images, both linear and kernel-based (nonlinear), may further enhance change signals relative to no-change background. IDL (Interactive Data Language) implementations of IR-MAD, automatic radiometric normalization and kernel PCA/MAF/MNF transformations are presented which function as transparent and fully integrated extensions of the ENVI remote sensing image analysis environment. Matlab code is also available which allows for fast data xploration and experimentation with smaller datasets. New, multi-resolution versions of IR-MAD which accelerate convergence and which further reduce no-change background noise are introduced. Computationally expensive matrix diagonalization and kernel image projections are programmed to run on massively parallel CUDA-enabled graphics processors, when available, giving an order of magnitude enhancement in computational speed. The software is available from the authors' websites.
KeywordsCUDA; ENVI; IDL; IR-MAD; iMAD; kPCA; kMAF; kMNF; Matlab; Radiometric Normalization; Remote Sensing; Multi-resolution
TypeJournal paper [With referee]
JournalComputers and Geosciences
Year2012    Vol. 38    No. 1    pp. 107-114
ISBN / ISSNdoi:10.1016/j.cageo.2011.05.012
NoteFor Matlab code see and For ENVI/IDL and Python code see Mort Canty's homepage.
Electronic version(s)[pdf]
BibTeX data [bibtex]
IMM Group(s)Geoinformatics