Road Junction Identification in High Resolution Urban SAR Images Based on SVM

Road extraction in high resolution urban SAR (Synthetic Aperture Radar) images is a difficult task because of multiplicative speckle noise and various interferences. Road junctions are important components of a road network. The quality of road extraction can be improved if road junctions are detected accurately. However, there are few related studies in this area. This study presents a road junction identification method with two stages. Firstly, global detection is performed to find the centre positions of the road junction candidates by using morphological operators. Secondly, the candidate road junctions are further identified based on the SVM (Support Vector Machine) classifier. The feature vectors used for SVM classification are Zernike moments. The proposed method is validated by two different SAR images. The results indicate that the proposed method has a higher identification accuracy than two other feature based SVM methods.