드론용 소형 SAR 시스템의 GPU 기반 실시간 영상복원 연구

Abstract

Although the back-projection algorithm(BPA) is the best method for image reconstruction of air-borne synthetic-aperture radar(SAR), it requires a long processing time. In this study, a real-time processing method of BPA was proposed to reconstruct the image of a compact drone-borne SAR in real time. To achieve this, the raw data of the SAR was acquired by a test flight with SAR based on frequency-modulated continuous wave(FMCW) radar. The performance of the BPA was confirmed by using raw data and a modified BPA processing kernel based on graphics processing unit(GPU) technology. The results confirmed the possibility of real-time image reconstruction of SAR.