Filtering of Polarimetric SAR Imagery Based on Multiplicative Model

A class of filters for polarimetric SAR imagery is proposed in this paper.It is assumed that polarimetric SAR data include two kind of information,texture information and polarimetric information,of which both are independent.In the multiplicative model,polarimetric covariance matrix is the carrier of those kinds of information expressed by total power and normalized covariance matrix and the original covariance matrix is decomposed to total power and normalized polarimetric covariance matrix.For the filters proposed in this study,texture and polarimetric information are separated firstly,and then filtered respectively,and finally covariance matrix is generated based on the multiplicative model.Experiments demonstrate that the proposed filter is valid and effective and polarimetric information is preserved well.