Robust Endmember Extraction in the Presence of Anomalies
Most available methods for end-member extraction use the convexity of the data structure and consider the vertices of the data as the purest pixels. Such methods do not consider the applicability of the linear mixing model once the end-members have been extracted. Thus they might return false end-members if the data contain outliers such as anomalies. In this paper, the authors tackle this problem by identifying end-members in a robust way, separating them from outliers. They tested the proposed algorithm with real and synthetic data and compared it with the VCA, SGA and N-FINDR algorithms, showing better and more robust end-member extraction.