An automated matching of real and simulated SAR image

Tie-points identification is much difficult for mountainous areas because of the huge different appearance of the SAR image and the map as well as the speckle noise present in the SAR image. However, tie-points under mountainous areas are very useful in many applications. Based on this requirement, this paper proposes an automated matching of a real and a simulated SAR image for control points (CP) selection. The method involves an image segmentation-based and clustering-based selection of the most distinctive CPs in simulated SAR image. The search for the corresponding CPs in the real SAR image is based on local similarity by means of template matching. To improve the precision, a method for segmentation threshold calculation and outlier screening is presented. Finally, the real SAR image is registered to the simulated one by using the polynomial mapping function. Experiment results on real data sets demonstrate the accuracy and efficiency of the proposed algorithm.