关键词: 反无人机检测, 反无人机数据集, 可见光, 红外, 射频, 雷达, 声学

Abstract: With the rapid proliferation of low-altitude unmanned aerial vehicles (UAVs), safety risks caused by illegal activities such as unauthorized and reckless flights have become increasingly prominent, highlighting an urgent need for efficient and reliable anti-UAV detection technologies. UAVs possess intrinsic characteristics that distinguish them from natural background objects in terms of appearance configuration, motion mechanisms, and communication links. These characteristics manifest in the observation space as static appearance differences, dynamic temporal differences, and intrinsic signal differences, which together constitute the fundamental basis for anti-UAV detection.From the perspective of UAV detectability based on intrinsic properties, this paper systematically reviews the development of existing anti-UAV detection technologies across three levels: static appearance differences, dynamic temporal differences, and intrinsic radio-frequency (RF) differences. Representative methods—including visible-light and infrared imaging, radar micro-Doppler analysis, acoustic rotor sound signatures, and communication-link emissions—are further summarized and analyzed with respect to their advantages and limitations under different sensing modalities and operational scenarios. In addition, publicly available datasets and commonly used evaluation metrics are compiled to facilitate fair and comprehensive comparative studies of related algorithms. This work aims to establish a hierarchical detection framework grounded in the intrinsic characteristics of UAVs and their contrast with background environments, thereby promoting the advancement of anti-UAV detection technologies.

Key words: Anti-UAV detection, Anti-UAV datasets, visible light, infrared, radio frequency (RF), radar, acoustics