Title :
Relation between principal components and endmembers in hyperspectral images
Author :
Blanco, D. ; Sanchez-Castillo, M. ; Carrión, M.C. ; Tienda-Luna, I.M.
Author_Institution :
Dept. of Appl. Phys., Univ. of Granada, Granada, Spain
Abstract :
In this contribution, the relation between the principal components of the covariance matrix of a hyperspectral image and the spectra of the endmembers is studied. When the data satisfy the spectral mixing model, from this relation the spectra of the endmembers and the abundance of each endmember in the pixels of the image can be theoretically obtained through a non-lineal minimization process. The simple case of an scene with two endmembers is studied using simulations.
Keywords :
covariance matrices; geophysical image processing; principal component analysis; covariance matrix; endmembers analysis; hyperspectral images; non-lineal minimization process; principal component analysis; spectral mixing model; Cost function; Covariance matrix; Equations; Estimation; Hyperspectral imaging; Matrix decomposition; Noise; Endmembers Analysis; Hyperspectral images; Principal Component Analysis;
Conference_Titel :
Geoscience and Remote Sensing Symposium (IGARSS), 2011 IEEE International
Conference_Location :
Vancouver, BC
Print_ISBN :
978-1-4577-1003-2
DOI :
10.1109/IGARSS.2011.6049467
