Orthogonal Subspace Projection Target Detector for Hyperspectral Anomaly Detection

Orthogonal subspace projection (OSP) is a versatile hyperspectral imaging technique which has shown great potential in dimensionality reduction, target detection, spectral unmixing, etc. However, due to its inherent requirement of prior knowledge, OSP not been explored anomaly detection. This article takes advantage an unsupervised OSP-based algorithm, automatic generation process (ATGP), and recently developed OSP-go decomposition (OSP-GoDec) along with data sphering (DS) make applicable Its idea implement ATGP on the background (BKG) subspaces constructed from low-rank matrix L sparse S generated by OSP-GoDec derive detector (OSP-AD). In particular, OSP-AD also includes DS remove BKG interference so as enhance Surprisingly, operating samples different constructions yields various versions OSP-AD. Experiments show that given appropriate construction subspace, can be outperform existing detectors including Reed-Xiaoli collaborative representation-based (CRD).