Abstract:

It is extremely difficult to extract road networks from high resolution synthetic aperture radar (SAR) images due to the presence of various disturbances. Markov random field (MRF) model can make full use of the imagery contextual characters and priori knowledge, which have been widely used to extract road networks. However, there exist some problems such as slow solution and many parameters setting of these type methods. In order to reduce the computation of subsequent iterative solution of MRF, pre-linking is firstly introduced to remove numerous false line elements based on the spatial relationship among them. Then the improved road networks Markov function model is established to label road networks. SAR images with 1 meter resolution are tested in the experiment. The results show the effectivity of the method mentioned above in high resolution SAR imagery road network extraction.