A Novel Implementation of a Hyperspectral Anomaly Detection Algorithm For Real Time Applications With Pushbroom Sensors

Anomaly detection is an increasingly important task when dealing with hyperspectral images in order to distinguish rare objects whose spectral characteristics substantially deviates from those of the neighboring materials. In this paper, a novel technique for accurate detection of anomalies in hyperspectral images is introduced. One of the main features of this method is its ability to process pushbroom data on-the-fly (i.e., line-by-line), being clearly suitable for real time applications in which memory resources are restricted as there is no need to store the whole hypercube. Diverse quality metrics have been applied on testing with real and synthetic hyperspectral data sets in order to compare the accuracy of the proposed algorithm over the state-of-the-art, showing the goodness of our proposal.