首页|基于深度学习的无人机检测和识别研究综述

基于深度学习的无人机检测和识别研究综述

扫码查看
原文链接
  • NETL
  • NSTL
  • 万方数据
近年来,由于在各行各业发挥了不可替代作用,无人机产业和应用得到了迅速发展.然而,无人机的"黑飞"、携带危险物品等事件频繁发生,对社会安全构成了严重威胁.因此,无人机的检测和识别变得尤为迫切和必要.随着无人机类型不断地变化,传统的检测与识别方法已不再适应当前需求.深度学习技术的快速发展为无人机检测与识别提供了一种高效且准确的解决方案.深度学习模型具备自主学习特征的能力,能够从大规模数据中提取高级特征,并且在无人机检测与识别任务中表现出色.该模型不仅能够显著提高准确性,还能够适应各种复杂环境和无人机类型.对此,本文综述了基于深度学习的无人机检测与识别技术的最新进展,主要包括基于深度学习的无人机视觉检测和识别、基于深度学习的无人机音频检测和识别、基于深度学习的无人机雷达检测和识别以及基于深度学习的无人机射频检测和识别.最后,对目前无人机检测和识别现存问题进行分析,并展望了未来研究方向.
Survey on UAV Detection and Identification Based on Deep Learning
In recent years,the unmanned aerial vehicle(UAV)industry and its applications have witnessed rapid devel-opment due to their irreplaceable roles in various sectors.However,incidents such as"black flights"of UAVs and those involving the carriage of dangerous goods occur frequently.Therefore,the detection and recognition of UAVs have be-come imperative.Traditional UAV detection and identification methods,relying mainly on rule-based and classical com-puter vision approaches,have become increasingly inadequate.and unable to meet the demands of contemporary secu-rity needs.It is within this context that the rapid advancement of deep learning technology has emerged as a game-changer,offering a potent and precise solution for UAV detection and recognition.Deep learning models,characterized by their remarkable ability to autonomously learn intricate features and extract high-level representations from vast and complex datasets,have demonstrated exceptional efficacy in the realm of UAV detection and identification.These mod-els have not only substantially elevated levels of accuracy but have also showcased adaptability to the vast spectrum of complex environmental conditions and UAV types.The incorporation of deep learning techniques into UAV security sys-tems has ushered in a new era of precision and efficiency.This comprehensive review paper seeks to provide a detailed exposition of the latest advancements in UAV detection and recognition technology grounded in deep learning principles.The paper explores a spectrum of modalities,including UAV visual detection and identification,UAV audio detection and identification,UAV radar detection and identification,and UAV radio frequency detection and identification.Each of these modalities represents a unique dimension in enhancing the overall effectiveness of UAV security systems.In con-clusion,this paper concludes by conducting a meticulous analysis of the current issues pervading the field of UAV detec-tion and recognition.Additionally,it offers a forward-looking perspective on the future research directions that hold the potential to further fortify the security and efficiency of UAV applications in an ever-changing technological landscape.

unmanned aerial vehicle(UAV)detection and recognitiondeep learningradio frequency

那振宇、程留洋、孙鸿晨、林彬

展开 >

大连海事大学信息科学技术学院,辽宁大连 116026

广东省空天通信与网络技术重点实验室,广东深圳 518055

无人机 检测和识别 深度学习 射频

国家自然科学基金国家自然科学基金国家自然科学基金辽宁省自然科学基金辽宁省"兴辽英才计划"项目广东省空天通信与网络技术重点实验室项目(2023)中央高校基本科研业务费专项

6197108161971083623710852023-MS-124XLYC20020783132023514

2024

10.16798/j.issn.1003-0530.2024.04.001

信号处理
中国电子学会

信号处理

CSTPCD北大核心
影响因子：1.502
ISSN：1003-0530
年,卷(期)：2024.40(4)
  • 6