Cargando…

Lightweight Object Detection Algorithm for UAV Aerial Imagery

Addressing the challenges of low detection precision and excessive parameter volume presented by the high resolution, significant scale variations, and complex backgrounds in UAV aerial imagery, this paper introduces MFP-YOLO, a lightweight detection algorithm based on YOLOv5s. Initially, a multipat...

Descripción completa

Detalles Bibliográficos
Autores principales: Wang, Jian, Zhang, Fei, Zhang, Yuesong, Liu, Yahui, Cheng, Ting
Formato: Online Artículo Texto
Lenguaje:English
Publicado: MDPI 2023
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10346989/
https://www.ncbi.nlm.nih.gov/pubmed/37447639
http://dx.doi.org/10.3390/s23135786
_version_ 1785073443463495680
author Wang, Jian
Zhang, Fei
Zhang, Yuesong
Liu, Yahui
Cheng, Ting
author_facet Wang, Jian
Zhang, Fei
Zhang, Yuesong
Liu, Yahui
Cheng, Ting
author_sort Wang, Jian
collection PubMed
description Addressing the challenges of low detection precision and excessive parameter volume presented by the high resolution, significant scale variations, and complex backgrounds in UAV aerial imagery, this paper introduces MFP-YOLO, a lightweight detection algorithm based on YOLOv5s. Initially, a multipath inverse residual module is designed, and an attention mechanism is incorporated to manage the issues associated with significant scale variations and abundant interference from complex backgrounds. Then, parallel deconvolutional spatial pyramid pooling is employed to extract scale-specific information, enhancing multi-scale target detection. Furthermore, the Focal-EIoU loss function is utilized to augment the model’s focus on high-quality samples, consequently improving training stability and detection accuracy. Finally, a lightweight decoupled head replaces the original model’s detection head, accelerating network convergence speed and enhancing detection precision. Experimental results demonstrate that MFP-YOLO improved the mAP50 on the VisDrone 2019 validation and test sets by 12.9% and 8.0%, respectively, compared to the original YOLOv5s. At the same time, the model’s parameter volume and weight size were reduced by 79.2% and 73.7%, respectively, indicating that MFP-YOLO outperforms other mainstream algorithms in UAV aerial imagery detection tasks.
format Online
Article
Text
id pubmed-10346989
institution National Center for Biotechnology Information
language English
publishDate 2023
publisher MDPI
record_format MEDLINE/PubMed
spelling pubmed-103469892023-07-15 Lightweight Object Detection Algorithm for UAV Aerial Imagery Wang, Jian Zhang, Fei Zhang, Yuesong Liu, Yahui Cheng, Ting Sensors (Basel) Article Addressing the challenges of low detection precision and excessive parameter volume presented by the high resolution, significant scale variations, and complex backgrounds in UAV aerial imagery, this paper introduces MFP-YOLO, a lightweight detection algorithm based on YOLOv5s. Initially, a multipath inverse residual module is designed, and an attention mechanism is incorporated to manage the issues associated with significant scale variations and abundant interference from complex backgrounds. Then, parallel deconvolutional spatial pyramid pooling is employed to extract scale-specific information, enhancing multi-scale target detection. Furthermore, the Focal-EIoU loss function is utilized to augment the model’s focus on high-quality samples, consequently improving training stability and detection accuracy. Finally, a lightweight decoupled head replaces the original model’s detection head, accelerating network convergence speed and enhancing detection precision. Experimental results demonstrate that MFP-YOLO improved the mAP50 on the VisDrone 2019 validation and test sets by 12.9% and 8.0%, respectively, compared to the original YOLOv5s. At the same time, the model’s parameter volume and weight size were reduced by 79.2% and 73.7%, respectively, indicating that MFP-YOLO outperforms other mainstream algorithms in UAV aerial imagery detection tasks. MDPI 2023-06-21 /pmc/articles/PMC10346989/ /pubmed/37447639 http://dx.doi.org/10.3390/s23135786 Text en © 2023 by the authors. https://creativecommons.org/licenses/by/4.0/Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
spellingShingle Article
Wang, Jian
Zhang, Fei
Zhang, Yuesong
Liu, Yahui
Cheng, Ting
Lightweight Object Detection Algorithm for UAV Aerial Imagery
title Lightweight Object Detection Algorithm for UAV Aerial Imagery
title_full Lightweight Object Detection Algorithm for UAV Aerial Imagery
title_fullStr Lightweight Object Detection Algorithm for UAV Aerial Imagery
title_full_unstemmed Lightweight Object Detection Algorithm for UAV Aerial Imagery
title_short Lightweight Object Detection Algorithm for UAV Aerial Imagery
title_sort lightweight object detection algorithm for uav aerial imagery
topic Article
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10346989/
https://www.ncbi.nlm.nih.gov/pubmed/37447639
http://dx.doi.org/10.3390/s23135786
work_keys_str_mv AT wangjian lightweightobjectdetectionalgorithmforuavaerialimagery
AT zhangfei lightweightobjectdetectionalgorithmforuavaerialimagery
AT zhangyuesong lightweightobjectdetectionalgorithmforuavaerialimagery
AT liuyahui lightweightobjectdetectionalgorithmforuavaerialimagery
AT chengting lightweightobjectdetectionalgorithmforuavaerialimagery