Moving Target Tracking for Single UAV In Open Outdoor Environment

The task of tracking a moving target by the UAV has two major challenges: one is that the multi-angle and multi-scale changes caused by the target motion make the target detection more difficult, and the other is the target loss due to variable direction of movement. Most of the existing research is...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:2020 6th International Conference on Big Data and Information Analytics (BigDIA) s. 317 - 324
Hlavní autoři: Xiao, Zhenliang, Liu, Daqian, Fei, Bowen, Men, Tong, Zhou, Zhijian, Zhang, Xiongtao, Zhou, Yun
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: IEEE 01.12.2020
Témata:
Data models
Mass Spring Algorithm (MSA)
Moving Target
Object detection
Open Outdoor Environment
Position measurement
Springs
Target tracking
Task analysis
Tracking Strategy
UAV
Vision sensors
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:The task of tracking a moving target by the UAV has two major challenges: one is that the multi-angle and multi-scale changes caused by the target motion make the target detection more difficult, and the other is the target loss due to variable direction of movement. Most of the existing research is limited to simulation studies, while the remaining few studies only allow the UAV to track specific target in specific environments. In order to solve these problems, we design and validate a new system framework for UAV tracking moving target in an open outdoor environment. In this framework, we use YOLOv3 to detect the multi-angle and multi-scale moving target, locate the target based on monocular vision, and design a tracking strategy, Mass Spring Algorithm (MSA), to prevent the loss of moving target. The experimental video 1 1 https://www.bilibili.com/video/BV1SD4y197gF/ shows that the system framework we designed is effective and successful. The results of comparative experiments to MSA shows that it effectively prevents the loss of target in the tracking process.
DOI:10.1109/BigDIA51454.2020.00057