Automated Extraction Of Control Points Based On SAR Image Simulation

A Synthetic Aperture Radar (SAR) image is simulated by Digital Elevation Model data, and the matching between simulated image and real image to select control points is the effective approach, which can come true SAR image automation in calculating orientation parameter. In this thesis the emphaseses are the SAR image simulation and SAR image matching. The main work and innovations of the thesis can be summerized as follows:1. The basic theory of SAR image simulation is introduced, it mainly includes SAR image'digital imaging theory, imaging model, geometric characters and radiometric characters.2. The current SAR image simulation method is analyzed and summerized, one is coherent simulation method, and the other is incoherent simulation method, and a new method is put forward based on the latter. The facet model is imported into the latter method, which makes the silmulation effect more reality.3. Using the ERS experimental data, the experiment of SAR image simulation is made aiming at hilly area. The simulated image is same with the real image, and has better textures.4. Due to the difference between simulated image and real image, the Scale Invariant Feature Transform (SIFT) arithmetic is adopted, which has invariability, and a new image bidirectional matching tactic is presented. Through the experiment,it rejects the unstable match feature points, and has better matching effect.