UAV target tracking method based on global feature interaction and anchor-frame-free perceptual feature modulation.

Target tracking techniques in the UAV perspective utilize UAV cameras to capture video streams and identify and track specific targets in real-time. Deep learning UAV target tracking methods based on the Siamese family have achieved significant results but still face challenges regarding accuracy an...

Full description

Saved in:
Bibliographic Details
Main Authors: Yuanhong Dan, Jinyan Li, Yu Jin, Yong Ji, Zhihao Wang, Dong Cheng
Format: Article
Language:English
Published: Public Library of Science (PLoS) 2025-01-01
Series:PLoS ONE
Online Access:https://doi.org/10.1371/journal.pone.0314485
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1832540274527567872
author Yuanhong Dan
Jinyan Li
Yu Jin
Yong Ji
Zhihao Wang
Dong Cheng
author_facet Yuanhong Dan
Jinyan Li
Yu Jin
Yong Ji
Zhihao Wang
Dong Cheng
author_sort Yuanhong Dan
collection DOAJ
description Target tracking techniques in the UAV perspective utilize UAV cameras to capture video streams and identify and track specific targets in real-time. Deep learning UAV target tracking methods based on the Siamese family have achieved significant results but still face challenges regarding accuracy and speed compatibility. In this study, in order to refine the feature representation and reduce the computational effort to improve the efficiency of the tracker, we perform feature fusion in deep inter-correlation operations and introduce a global attention mechanism to enhance the model's field of view range and feature refinement capability to improve the tracking performance for small targets. In addition, we design an anchor-free frame-aware feature modulation mechanism to reduce computation and generate high-quality anchors while optimizing the target frame refinement computation to improve the adaptability to target deformation motion. Comparison experiments with several popular algorithms on UAV tracking datasets, such as UAV123@10fps, UAV20L, and DTB70, show that the algorithm balances speed and accuracy. In order to verify the reliability of the algorithm, we built a physical experimental environment on the Jetson Orin Nano platform. We realized a real-time processing speed of 30 frames per second.
format Article
id doaj-art-b67232d7104544cc8e1f9a343010e0ba
institution Kabale University
issn 1932-6203
language English
publishDate 2025-01-01
publisher Public Library of Science (PLoS)
record_format Article
series PLoS ONE
spelling doaj-art-b67232d7104544cc8e1f9a343010e0ba2025-02-05T05:31:21ZengPublic Library of Science (PLoS)PLoS ONE1932-62032025-01-01201e031448510.1371/journal.pone.0314485UAV target tracking method based on global feature interaction and anchor-frame-free perceptual feature modulation.Yuanhong DanJinyan LiYu JinYong JiZhihao WangDong ChengTarget tracking techniques in the UAV perspective utilize UAV cameras to capture video streams and identify and track specific targets in real-time. Deep learning UAV target tracking methods based on the Siamese family have achieved significant results but still face challenges regarding accuracy and speed compatibility. In this study, in order to refine the feature representation and reduce the computational effort to improve the efficiency of the tracker, we perform feature fusion in deep inter-correlation operations and introduce a global attention mechanism to enhance the model's field of view range and feature refinement capability to improve the tracking performance for small targets. In addition, we design an anchor-free frame-aware feature modulation mechanism to reduce computation and generate high-quality anchors while optimizing the target frame refinement computation to improve the adaptability to target deformation motion. Comparison experiments with several popular algorithms on UAV tracking datasets, such as UAV123@10fps, UAV20L, and DTB70, show that the algorithm balances speed and accuracy. In order to verify the reliability of the algorithm, we built a physical experimental environment on the Jetson Orin Nano platform. We realized a real-time processing speed of 30 frames per second.https://doi.org/10.1371/journal.pone.0314485
spellingShingle Yuanhong Dan
Jinyan Li
Yu Jin
Yong Ji
Zhihao Wang
Dong Cheng
UAV target tracking method based on global feature interaction and anchor-frame-free perceptual feature modulation.
PLoS ONE
title UAV target tracking method based on global feature interaction and anchor-frame-free perceptual feature modulation.
title_full UAV target tracking method based on global feature interaction and anchor-frame-free perceptual feature modulation.
title_fullStr UAV target tracking method based on global feature interaction and anchor-frame-free perceptual feature modulation.
title_full_unstemmed UAV target tracking method based on global feature interaction and anchor-frame-free perceptual feature modulation.
title_short UAV target tracking method based on global feature interaction and anchor-frame-free perceptual feature modulation.
title_sort uav target tracking method based on global feature interaction and anchor frame free perceptual feature modulation
url https://doi.org/10.1371/journal.pone.0314485
work_keys_str_mv AT yuanhongdan uavtargettrackingmethodbasedonglobalfeatureinteractionandanchorframefreeperceptualfeaturemodulation
AT jinyanli uavtargettrackingmethodbasedonglobalfeatureinteractionandanchorframefreeperceptualfeaturemodulation
AT yujin uavtargettrackingmethodbasedonglobalfeatureinteractionandanchorframefreeperceptualfeaturemodulation
AT yongji uavtargettrackingmethodbasedonglobalfeatureinteractionandanchorframefreeperceptualfeaturemodulation
AT zhihaowang uavtargettrackingmethodbasedonglobalfeatureinteractionandanchorframefreeperceptualfeaturemodulation
AT dongcheng uavtargettrackingmethodbasedonglobalfeatureinteractionandanchorframefreeperceptualfeaturemodulation