视觉感知反无人机技术发展动态与趋势  

Visual perception-based anti-drone technology:Development dynamics and trends

作　　者：赵芬 赵长春 郭俊 杨俊波[1] ZHAO Fen;ZHAO Changchun;GUO Jun;YANG Junbo(College of Science,National University of Defense Technology,Changsha 410073,China;School of Artificial Intelligence,Chongqing University of Technology,Chongqing 400054,China)

机构地区：[1]国防科技大学理学院,湖南长沙410073 [2]重庆理工大学两江人工智能学院,重庆400054

出　　处：《国防科技》2023年第5期35-45,共11页National Defense Technology

基　　金：国家重点研发计划(2022YFF0706005);国家自然科学基金项目(12272407、60907003、61805278、62275269、62275271)。

摘　　要：在深度学习持续发展的背景下,视觉感知目标检测技术取得了显著的发展,利用计算机视觉和图像处理技术对无人机目标进行识别与跟踪,进而对无人机目标的行为趋势进行预测,可以极大提高无人机目标检测的精度和速度,对于保障国防安全具有重要的研究意义和应用价值。首先对国内外不同类别的反无人机技术进行阐述;其次从传统目标检测方法和基于深度学习的目标检测方法两个方面分别介绍相关技术与核心算法;最后对未来视觉感知反无人机技术的发展趋势和存在的问题进行分析,并展望反无人机领域的未来发展前景。With the continuous advancement of deep learning,significant progress has been made in visual perception-based object detection technology.The use of visual perception technology can greatly enhance the accuracy and speed of drone object detection.Different categories of anti-drone technology from domestic and international perspectives are elaborated,followed by an introduction to traditional and deep-learning-based detection techniques and core algorithms.Existing visual-perception-based anti-drone technology exhibits high timeliness and accuracy,enabling the real-time and accurate detection and tracking of drone targets.However,challenges remain in the detection and tracking of drones in complex environments,of low,slow,and small drones,and of hidden drones.Looking into the future,anti-drone technology holds vast potential in areas such as multimodal fusion,automated decision-making,and predictive analysis of drone behavior.Related technologies will provide robust support to the field of anti-drone technology,enhancing perception and response capabilities to drone threats.

关 键 词：反无人机 目标检测 视觉感知 深度学习 

分 类 号：V279[航空宇航科学与技术—飞行器设计] TP391.41[自动化与计算机技术—计算机应用技术]