Research status of hyperspectral anomaly target detection

Abstract

Hyperspectral imagery is a three-dimensional image of “image-spectrum merging”. Its spectral resolution reaches nanometer level, which provides a theoretical basis for classification, unmixing and target detection of features. Because the anomaly target detection does not need to obtain the prior knowledge of features, it is more widely used in practice and has become a research hotspot. This paper summarizes the development status of anomaly target detection algorithms at home and abroad, and expounds the existing problems. Finally, the development trend of anomaly target detection is prospected.