Unmixing and anomaly detection in hyperspectral data dueto cluster variation and local information

摘要

This paper presents a novel method for anomaly detection based on a cluster unmixing approach. Severalalgorithms for endmember extraction and unmixing have been reported in literature. Endmember extractionalgorithms search for pure materials which constitute the significant structure of the environment. For abundanceestimation in hyperspectral imagery, various physically motivated least squares methods are considered. In realhyperspectral data, signatures of each pure material vary with physical texture and perspective. In this work,clustering of data is performed and normal distributions - instead of constant signatures - are used to represent theendmembers. This representation allows determination of class membership by means of unmixing. Furthermore,a parameter optimization is performed. Using only endmembers in a focal window around each pixel better fitsthe physical model. As result of this local approach, the residual of the reconstruction indicates the magnitudeof anomalies. The results obtained with the new approach is called 'Cluster Mixing' (CM). The performance ofCluster Mixing is illustrated by a comparison with other anomaly detection algorithms.