Simplex ACE: a constrained subspace detector

Physics-guided neural networks for hyperspectral target identification

“…Hyperspectral target detection is the process of determining if any hyperspectral image pixels contain specific solid materials or gases (the "targets") by comparing each pixel-under-test against the reference spectra of the target(s); 10,11 standard detection methods include matched filters 12 and variants of the adaptive matched filter (AMF) [13][14][15][16][17][18][19][20][21][22] and adaptive cosine/coherence estimator (ACE). [23][24][25][26][27][28] For large spectral libraries, the sensing problem is sometimes framed as a two-step process consisting of target detection and identification; we note that spectral library sizes often inflate significantly when including multiple representations of targets with high variability in their spectra (e.g., due to morphology). [29][30][31][32] In a combined detection and identification procedure, candidate target pixels are first detected, and then target identification traditionally consists of comparing a region of interest (ROI) consistent of detected pixels against the spectral library, 33,34 often using tools such as linear regression.…”

Hyperspectral Target Detection for Multiple Target Spectra

“…A review of target detection algorithms is discussed in [1]. The conventional detectors for target detection are the Spectral Angle Mapper (SAM), Adaptive Matched Filter (AMF) [2], and the Adaptive Coherence/Cosine Angle [3].…”