YOLO 기반의 SAR을 이용한 지상 이동 표적 탐지 기법

초록

Synthetic aperture radar (SAR)-based moving target detection plays a crucial role in fields requiring real-time surveillance because it enables moving target detection regardless of weather conditions or time of day. Widely used techniques such as displaced phase center antennas or along-track interferometry produce numerous false alarms when using constant false alarm rate detectors for moving target detection. This study proposed a you only look once-based moving-target detection method that uses phase maps from multiple SAR images as inputs. The performance of the proposed method was validated through simulations under various signal-to-clutter and noise (SCNR) conditions. The proposed method improved the detection rate by more than 15 % compared to conventional methods at an SCNR of 5 dB while maintaining a detection rate of more than 90 % at SCNR levels below 0 dB, where conventional methods fail to detect targets.

키워드