Detection of Targets in Polarimetric Radar Images Using Orthogonal Subspace Projection

Detection and characterization of the constituent scatterers for each pixel in the scene is one of the fundamental goals of microwave remote sensing. Compared with synthetic aperture radar (SAR) sensing, polarimetric SAR (PolSAR) sensing has finer details of the scattering properties of targets and natural background; therefore, it can increase the detectability of single and partial targets. In this paper, a new and fresh look at the target detection issue is taken and an effective technique which simultaneously annihilates interfering background is developed that detects the presence of a scattering mechanism of interest. Several canonical scattering mechanisms are assumed as our signal sources whose combination forms scattering vector of each pixel with appropriate weight fractions. Using this technique leads to the exact detection of partial targets and build-up areas.