Research On The Screening Of SAR Images For Typical Targets

As a microwave remote sensing means, Sythetic Aperture Radar (SAR) has eminent characteristics of all-time and all-weather, which has been widely applied in the fields of military and civilian. As the sensor technology progresses rapidly and the SAR data increases sharply, the interpretation of data is under more stress. So it is urgent to develop the technology of mass SAR image data screening. With the fast data screening technology ,the images ,which are unvalued and do not comprise interested targets, can be filtrated efficiently. The interpretation of SAR image data could become more easy and fast as a result of the screening technology.Around the mass SAR images screening technology, the thesis comprises the main work and the researching productions below:The SAR imaging theory has been studied thoroughly. Based on the pilot study of SAR image's general characteristics(such as the speckle,the geometry and the gray characteristics etc), The basic flow of SAR image interpretation is introduced.The SAR imaging characteristics of two typical targets(airplane and ship) have been analyzed in detail. The flow of feature extraction and target detection is described. In order to serve the image screening, the target clue and background characteristics have been analyzed. A group of SAR image gray histogram features is chosen to assist the image screening method.A SAR image screening method based on the significant bit-plane is proposed . After the suitable gray pretreatment, the significant bit-plane image is paid attention by the measurement of bit-plane complexity,runlength and frequency spectrum . In order to obtain the airport SAR images, combined with the gray histogram features, experiment shows that the method can screen the images fast. Besides, the airport target is extracted .The result can satisfy the requirement.The idea of statistic characteristic analysis is implemented in SAR image screening method. Combined with different statistic models and matching laws, the thesis conducts modeling analysis to SAR images of different scenes, then gets the optimum model and matching threshold. To obtain harbor or sea images ,experiment is carried out with the assist of image gray histogram features. The results prove the method's feasibility.