Share Email Print
cover

Proceedings Paper

Relationship between linear spectral unmixing and endmember finding
Author(s): Hsiao-Chi Li; Shih-Yu Chen; Cheng Gao; Chein-I Chang
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Linear spectral unmixing (LSU) and endmember finding are closely related. The link between these two has been recognized recently via their shared geometric structure, simplex. Since a fully abundance-constrained LSU (FAC-LSU) performs two physical constraints, Abundance Sum-to-one Constraint (ASC) and Abundance Non-negativity Constraint (ANC) it can be interpreted as the fact that all data sample vectors are embraced by a simplex with vertices formed by a set of signatures of interest used to unmix data sample vectors. In this case, all the data sample vectors inside the simplex are unmixed by their fully constrained abundance fractions, while all data sample vector outside the simplex are considered as unwanted sample vectors such as noisy samples, bad sample vectors. As for endmember finding it makes use of a simplex to locate and find endmembers which are assumed to be its vertices due to their signature purity and geometric convexity. In other words, in order to find endmembers finding a largest simplex that include as many data sample vectors as possible where the vertices of this simplex are considered endmembers. Such simplex is the one with maximal simplex volume. Interestingly, both LSU and endmember finding share one thing in common, which is their use of the same two constraints, ASC and ANC imposed by a simplex. This paper explores such relationship between LSU and endmember finding.

Paper Details

Date Published:
PDF: 14 pages
Proc. SPIE 9124, Satellite Data Compression, Communications, and Processing X, 91240M; doi: 10.1117/12.2049039
Show Author Affiliations
Hsiao-Chi Li, Univ. of Maryland, Baltimore County (United States)
Shih-Yu Chen, Univ. of Maryland, Baltimore County (United States)
Cheng Gao, Univ. of Maryland, Baltimore County (United States)
Chein-I Chang, Univ. of Maryland, Baltimore County (United States)


Published in SPIE Proceedings Vol. 9124:
Satellite Data Compression, Communications, and Processing X
Bormin Huang; Chein-I Chang; José Fco. López, Editor(s)

© SPIE. Terms of Use
Back to Top