Research And Implementation Of High Resolution SAR Image Battle Damage Assessment

With the development of military science and technology, BDA technique becomes more important in modern warfare. Great attention has been paid to this field by governments and researchers. Synthetic aperture radar meets the Real-time requirements of BDA because of its all-day, all-weather capabilities. So BDA based on SAR image is very important in military application.In this thesis, according to the defense pre-research fund, techniques of battle damage simulation and SAR image simulation of complicated region was studied. On the basis of analyzing the key techniques of image processing in BDA, we proposed a method for SAR image BDA, these algorithms are applied to achieve a SAR image BDA system. The main content of this thesis are summarized as follows:1. Research on battle damage simulation and SAR image simulation. The lack of after-strike images makes research on BDA harder. A algorithm was proposed to solve this problem. Through the number, shape, and intensity of fragments controlled by the fragment distribution model, fragments generation model and intensity change model, battle damage simulation was realized. Sometimes, there are no SAR images in designated region. According to the principle of SAR imagery and speckle theory, SAR image simulation method based on backscattering coefficient distribution of complicated region was studied to provide simulated SAR images.2. Some related image processing techniques about image-based BDA was studied. On the basis of battle damage simulation and SAR image simulation techniques, a research case about BDA system was designed. Then, some key techniques related to BDA: image registration, CFAR target detection, change detection, image segmentation using Otsu methods, are deeply analyzed and studied.3. SAR image BDA method based on change detection. The existing algorithm based on target feature comparison has more demanding on the target type. It greatly limits the application of this algorithm. A practical method was proposed. Segmented regions with different damage level were provided by using image registration, change detection, threshold segmentation.4. Design and realization of the SAR image BDA system. In this system, four modules including battle damage simulation, SAR image simulation, image registration, battle damage assessment are realized using Visual C++6.0 and Oracle /SQL Sever 9i.