首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于深度学习的目标跟踪技术及其多波段应用研究
作者姓名:李文涛  李世玉  郇金晓  李磊  张振华  康忱  刘娣  乔海东  牟进超
作者单位:西安电子科技大学天线微波技术国家重点实验室,西安电子科技大学天线微波技术国家重点实验室,西安电子科技大学天线微波技术国家重点实验室,西安电子科技大学天线微波技术国家重点实验室,北京遥测技术研究所,北京遥测技术研究所,北京遥测技术研究所,北京遥测技术研究所,北京遥测技术研究所
基金项目:中央军委173计划重点项目;北京市科技新星计划(Z201100006820130)
摘    要:目标跟踪是计算机视觉领域中最具挑战性的领域之一。首先,回顾了目标跟踪技术发展现状并总结出目标跟踪的一般流程;其次,分析了卡尔曼滤波、粒子滤波等传统目标跟踪方法的优缺点;再次,依次重点介绍了固定窗跟踪等相关滤波目标跟踪方法,以及全网络自适应目标跟踪等高性能目标跟踪方法;最后,总结了目标跟踪在红外、毫米波、太赫兹等波段的应用,并对多波段融合目标跟踪进行了展望。

关 键 词:目标跟踪  深度学习  太赫兹  红外  毫米波
收稿时间:2022/1/6 0:00:00
修稿时间:2022/11/8 0:00:00

Research on target tracking technology based on deep learning and its multi-band application
Authors:LI Wentao  LI Shiyu  HUAN Jinxiao  LI Lei  ZHANG Zhenhu  KANG Chen  LIU Di  QIAO Haidong and MOU Jinchao
Institution:State Key Laboratory of Antenna and Microwaves, Xidian University,State Key Laboratory of Antenna and Microwaves, Xidian University,State Key Laboratory of Antenna and Microwaves, Xidian University,State Key Laboratory of Antenna and Microwaves, Xidian University,Beijing Research Institute of Telemetry,Beijing Research Institute of Telemetry,Beijing Research Institute of Telemetry,Beijing Research Institute of Telemetry,Beijing Research Institute of Telemetry
Abstract:Target tracking is one of the most challenging topics in computer vision field. Firstly, the development of target tracking technology is reviewed and the general process of target tracking is summarized. Secondly, the advantages and disadvantages of traditional target tracking methods based on Kalman filter and particle filter are analyzed. Thirdly, the fixed window correlation filter target tracking methods and high-performance adaptive full network target tracking methods are successively highlighted. Finally, the applications of target tracking in infrared, millimeter wave and terahertz radiation band are summarized, and multi-band fusion target tracking is forecasted.
Keywords:Target tracking  Deep learning  Terahertz  Infrared  Millimeter wave
点击此处可从《》浏览原始摘要信息
点击此处可从《》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号