Urban area extraction in airborne SAR imagery based on the parameter of GOA distribution

An algorithm of extracting urban areas in airborne synthetic aperture radar (SAR) imagery is presented based on the local statistical characteristics. The G^O _A distribution is introduced to model the ground scene in SAR imagery firstly. Then we demonstrate that the parameter ¿ of G^O _A distribution in urban regions is much smaller than those containing non-artificial regions. Finally, experimental results using PiSAR data show that the proposed algorithm can distinguish urban areas from other areas better than the Weibull distribution based algorithm.