Target-to-Anomaly Conversion for Hyperspectral Anomaly Detection

A known target detection assumes that the to be detected is provided a priori , while an anomaly unknown without any prior knowledge. As a result, generally performs search-before-detect in active mode, referred as opposed which throw-before-detect passive detection. Accordingly, techniques designed for these two types of are completely different. This paper shows there indeed mechanism, called target-to-anomaly conversion can convert hyperspectral (HTD) (HAD) via novel idea, dummy variable trick (DVT). By virtue such many well-known techniques, likelihood ratio test (LRT), constrained energy minimization (CEM) and orthogonal subspace projection (OSP), spectral angle mapper (SAM) adaptive cosine estimator (ACE) converted their corresponding detector, conversion-derived detector (TAC-AD). Since requires knowledge TAC-AD does not, direct use RAC-AD not effective. To make work, newly developed approach effective space (EAS) implemented conjunction with so anomalies retained EAS interference noise including background (BKG) removed from EAS. The experiments demonstrate operating better than existing approaches model-based methods.