Automatic detection of military vehicles hidden among forests in synthetic aperture radar (SAR) image is a challenging and difficult task because tree trunk clutter often appears as locally bright as obscured targets. We apply subspace-based detection methods to treat the problem as detecting subspace signals in structured subspace interference and broadband noise of unknown level. Specifically, tree trunk clutter is modeled as structured isotropic interferences with unknown amplitudes while dominant vehicle scatterers as anisotropic dihedral responses. Matched subspace detector (MSD) with constant false alarm rate (CFAR) property is derived. Experiments on both simulated and real foliage-penetrating (FOPEN) SAR image show that the proposed detection scheme has good detection performance even at low false alarm rates.
Subspace-based CFAR detection of vehicles from forests in SAR image
International Conference on Space Information Technology ; 2005 ; Wuhan,China
Proc. SPIE ; 5985
2005-11-19
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Matched subspace CFAR detection of hovering helicopters
IEEE | 1999
|PAPERS - Matched Subspace CFAR Detection of Hovering Helicopters
Online Contents | 1999
|Online Contents | 1995
|IEEE | 1995
|