Research Progress of Multi-Target Tracking Based on Deep Learning from Perspective of UAV

被引:0
作者
Yang, Yang [1 ]
Song, Pinde [1 ]
Zhong, Chunlai [1 ]
Cao, Lijia [2 ,3 ,4 ]
机构
[1] School of Automation and Information Engineering, Sichuan University of Science & Engineering, Sichuan, Yibin
[2] School of Computing Science and Engineering, Sichuan University of Science & Engineering, Sichuan, Yibin
[3] Artificial Intelligence Key Laboratory of Sichuan Province, Sichuan, Yibin
[4] Key Laboratory of Higher Education of Sichuan Province for Enterprise Informationalization and Internet of Things, Sichuan, Yibin
关键词
deep learning; joint detection tracking; multi-target tracking; tracking by detection; unmanned aerial vehicle(UAV);
D O I
10.3778/j.issn.1002-8331.2307-0355
中图分类号
学科分类号
摘要
Multi-object tracking based on UAV platforms has wide application prospects in various fields, including smart cities, agricultural production, disaster early warning and search, and rescue operations. Unlike the relatively mature multi-object tracking from the traditional perspective, multi-object tracking from the UAV perspective faces a series of challenges that have not been completely solved. These challenges mainly include target scale changes, interference from similar targets, target occlusion and overlap, and uneven target distribution. This paper compiles the classic multiobject tracking algorithms developed in recent years from a traditional perspective. It also comprehensively analyzes the main technical approaches and latest methods in the field of multi-object tracking from the perspective of UAVs, with a particular focus on the detection-based tracking framework. Additionally, it examines the performance evaluation methods and mainstream datasets used in this domain. Moreover, the paper analyzes the primary challenges faced in multi-object tracking from UAV perspectives and offers insights into future research trends, aiming to provide valuable references for further related studies. © 2023 Chinese Medical Journals Publishing House Co.Ltd. All rights reserved.
引用
收藏
页码:48 / 62
页数:14
相关论文
共 84 条
[1]  
JIMENEZ-BRAVO D M, LOZANO M A, SALES MENDES A, Et al., Multi- object tracking in traffic environments:a systematic literature review[J], Neurocomputing, 494, pp. 43-55, (2022)
[2]  
LUO W,, XING J, MILAN A, Et al., Multiple object tracking:a literature review[J], Artificial Intelligence, 293, (2021)
[3]  
LIU Y, LI M M, ZHENG B Q,, Et al., Survey on video object tracking algorithms[J], Journal of Frontiers of Computer Science and Technology, 16, 7, pp. 1504-1515, (2022)
[4]  
YU F H, ZHANG H G, JIN Z Y,, Et al., Research status and prospect of unmanned aerial vehicle remote sensing in inversion of rice agronomic physicochemical parameters[J], Journal of Shenyang Agricultural University, 54, 2, pp. 248-256, (2023)
[5]  
ZHANG H R, SHEN L F,, REN Z Z,, Et al., A comprehensive survey on unmanned aerial vehicle assisted intelligent edge network[J], Telecommunication Engineering, 1, pp. 1-10, (2023)
[6]  
XU Y,, ZHOU X, CHEN S,, Et al., Deep learning for multiple object tracking:a survey[J], IET Computer Vision, 13, 4, pp. 355-368, (2019)
[7]  
PAL S K,, PRAMANIK A, Et al., Deep learning in multi-object detection and tracking:state of the art[J], Applied Intelligence, 51, 9, pp. 6400-6429, (2021)
[8]  
CHEN L, LIU Y G., UAV visual target tracking algorithms:review and future prospect[J], Information and Control, 51, 1, pp. 23-40, (2022)
[9]  
YUAN Y B, WU Y Q, ZHAO L Y,, Et al., Research progress of UAV aerial video multi-object detection and tracking based on deep learning[J], Acta Aeronautica et Astronautica Sinica, 18, pp. 1-31, (2023)
[10]  
GE B Y, ZUO X Z, HU Y J., Review of visual object tracking technology[J], Journal of Image and Graphics, 23, 8, pp. 1091-1107, (2018)