Research On The Spaceborne Synthetic Aperture Radar Target Detection

The thesis contacted closely with working characteristic and martial demand of satellite station of space-borne synthetic aperture radar (SAR). Based on analyzing and synthesizing typical algorithms of SAR target detection, the issue of SAR target detection is surveyed and studied. The thesis focuses on researching novel approaches and techniques with the purposes of real-time detection of selected targets and data handling. Beginning with researching algorithms of SAR imaging, we analyze SAR echo data, imagery and middle domain of SAR imaging process. Targets characters, possible target detected and potential applications are studied. New algorithms of target detection are developed with SAR images, range-doppler domain (RDD) and SAR echo data respectively.In the domain of SAR image, we analyze firstly the features of noise distribution and geometrical structure, review the presently advanced target detection techniques and gave experimental results with real SAR imagery. Then extraction of distinct structures from SAR image are studied, and a detection method of multi-scale features is developed ,which is called as Topographical Primal Sketch (TPS). Based on it, a new algorithm is presented for detecting point-set targets.In the domain of RDD, target characters in RDD image are theoretically analyzed, and description and extraction of target features are studied. A new classification method is presented based on processing and statistic analysis with a mount of ERS-1/2 and Radarsat-1 data. And an approach to coastline extraction is followed to be developed, which uses both RDD and sub-aperture correlation features. Also approach adaptability to case of squint SAR system is discussed.In the domain of echo data, two different technical ways to solve the problem of moving target detection and two relative new approaches are represented based on analysis of target characters. One is MCS/MTD approach, which is modifying phase compensation factor in the steps of chirp-scaling (CS) algorithm, and the other is Dechirp-MTD approach, which is based on De-chirp imaging technique.All algorithm presented are applied to real space-borne SAR data, and the experimental results show their validity and adaptability.