SAR IMAGE SIMULATION WITH APPLICATION TO TARGET RECOGNITION

This paper presents a novel synthetic aperture radar (SAR) image simulation approach to target recognition, which consists of two frameworks, referred to as the satellite SAR images simulation and the target recognition and identiflcation. The images simulation makes use of the sensor and target geo-location relative to the Earth, movement of SAR sensor, SAR system parameters, radiometric and geometric characteristics of the target, and target radar cross section (RCS), orbital parameters estimation, SAR echo signal generation and image focusing to build SAR image database. A hybrid algorithm that combines the physical optics, physical difiraction theory, and shooting and bouncing rays was used to compute the RCS of complex radar targets. Such database is vital for aided target recognition and identiflcation system Followed by reformulating the projection kernel in an optimization equation form, the target's re∞ectivity fleld can be accurately estimated. Accordingly, the target's features can be efiectively enhanced and extracted, and the dominant scattering centers are well separated. Experimental results demonstrate that the simulated database developed in this paper is well suited for target recognition. Performance is extensively tested and evaluated from real images by Radarsat-2 and TerraSAR-X. Efiectiveness and e-ciency of the proposed method are further conflrmed.