Automatic extraction of power transmission tower series from PolSAR imagery based on MRF model

This paper describes a three-step algorithm for automatic extraction of power transmission tower series in full polarimetric SAR imagery. Firstly, the method uses a polarimetric whitening filter to combine the information in different polarimetric channels and reduces the speckle. Then the point-like targets are detected by an adaptive region classification CFAR detection approach, among which only a few are the potential power transmission towers. Finally, a graph linked the potential point targets is constructed and a Markov random field is defined on the constructed graph, the extraction result is obtained through an energy minimization process. The experimental result verifies the effectiveness of this algorithm.