Title :
A least-squares-filter vector hybrid approach to hyperspectral subpixel demixing
Author_Institution :
Sci. Applications Int. Corp., Arlington, VA, USA
fDate :
3/1/1999 12:00:00 AM
Abstract :
The least-squares and filter vector algorithm (FVA) approaches to subpixel demixing are mathematically equivalent. This equivalence is used to construct a hybrid approach to demixing. This approach maintains the speed of the filter vector algorithm while still providing the least-squares expressions for coefficient variances, weighed fits, and constrained fits
Keywords :
geophysical signal processing; geophysical techniques; least squares approximations; multidimensional signal processing; remote sensing; terrain mapping; data model; filter vector algorithm; geophysical measurement technique; hybrid approach; hyperspectral remote sensing; hyperspectral subpixel demixing; land surface; least-squares; least-squares-filter vector hybrid approach; multidimensional signal processing; multispectral remote sensing; optical imaging; subpixel demixing; terrain mapping; Chemicals; Data models; Equations; Filters; Hyperspectral imaging; Hyperspectral sensors; Least squares methods; Linear regression; Object detection; Vectors;
Journal_Title :
Geoscience and Remote Sensing, IEEE Transactions on
