Polarimetric SAR image change detection based on three-component decomposition

On the basis of the polarimetric distance measurement, this paper presents an improved method of change detection through a complete model-based decomposition of the polarimetric synthetic aperture radar (POLSAR) covariance/coherency matrix data. Due to the capability of accurately characterizing the structural features for polarization decomposition, the proposed method first utilizes the Freeman-Durden decomposition (FDD) to extract the polarization scattering characteristics of different land objects. Then, the refined Lee filter is applied to reduce the speckle noise. Finally, an improved distance measurement of change detection for multi-temporal POLSAR data is proposed. ALOS PALSAR polarimetric data from October 08, 2008 and November 23, 2008 covering agricultural fields near north Toronto, Canada are used. The experiments demonstrate that the proposed method can distinguish the changes effectively and preferably compared with the conventional method based on the distance measurement.