Polarimetric SAR image fusion using nonnegative matrix factorisation and improved-RGB model

Abstract

Since different polarimetrie channels have different dynamical ranges, nonnegative matrix factorisation (NMF) is explored for fusing polarimetric SAR (PolSAR) data in a single colour image. The data of the three polarimetric channels (HH, HV and VV) with total power (TP) data is fully utilised to implement the NMF, and the choice of normalised factors for different polarimetric channels based on the same probability is applied, which produces an improved-RGB model. Combining the improved-RGB model with NMF, this approach brings in fusion images with higher spatial resolution and the power component preserved.