合成孔径雷达
计算机科学
稳健性(进化)
鉴定(生物学)
逆合成孔径雷达
雷达
可扩展性
雷达成像
计算机视觉
遥感
人工智能
实时计算
电信
地质学
生物化学
化学
植物
数据库
生物
基因
标识
DOI:10.1109/igarss52108.2023.10282026
摘要
Military target identification is critical for defense and intelligence operations, but acquiring labeled Synthetic Aperture Radar (SAR) data poses challenges. This paper introduces an innovative approach that leverages SAR simulation to generate labeled data, addressing security constraints, operational limitations, and high costs associated with data collection. The integration of both phase and amplitude images obtained from SAR simulation was investigated to enhance target identification accuracy and robustness. SAR simulation offers a cost-effective and scalable solution, accurately modeling the radar imaging process to create large quantities of labeled data resembling real-world scenarios. The Cross-Track Interferometry SAR (XTI-SAR), utilizing a vertical baseline in the direction of platform flight, was employed to convert phase information into target height. The outcomes contribute to advancing SAR-based target identification techniques, providing valuable insights for effective and efficient target identification in complex operational scenarios.
科研通智能强力驱动
Strongly Powered by AbleSci AI