Abstract:

To improve the accuracy of high resolution polarimetric synthetic aperture radar (SAR) four-component decomposition, an adaptive coherent matrix estimation method is proposed, which is applied before decomposition. First, in order to eliminate the effect of the extreme heterogeneity cell, proper samples are chosen adaptively according to the main types of cell scattering. Then, the ensemble average estimate and the fixed point estimate are applied separately according to the degree of feature heterogeneous to accommodate different feature distribution. This method is validated by using the ESAR data and the UAVSAR data. The results show that the decomposition results are closer to real scattering types by applying the adaptive matrix estimation.