Using spectral-shape parameters to improve linear spectral mixture analysis
Linear spectral mixture analysis (LSMA) has been frequently used to derive sub-pixel information from moderate-resolution satellite images. This letter proposes a new method to improve LSMA using spectral-shape parameters. A Mann-Whitney U test, Wilcoxon W test statistical analysis and root mean square error (RMSE) were used to compare the fractions estimated from satellite images using traditional LSMA with a 'shade' endmember (LSMAWS), the normalized spectral mixture by mean ratio (NSMMR), the proposed spectral-shape-based LSMA (SSLSMA) and the 'actual' fractions generated from an ortho-image quarter quadrangle. These statistical analyses suggest that the accuracy was significantly improved using the spectral-shape-based LSMA model in identifying landscape classes at the sub-pixel level.
No Reference information available - sign in for access.
No Citation information available - sign in for access.
No Supplementary Data.
No Article Media
Document Type: Research Article
Affiliations: Department of Geography and Environmental Systems, University of Maryland, Baltimore County, 1000 Hilltop Circle, Baltimore, MD, USA
Publication date: 01 January 2009