Port terminal mobile recognition based on combined YOLOv5s-DeepSort.

To solve the problem of reduced positioning accuracy caused by changes in scale, background and occlusion in port and dock video images, this research proposes an enhanced model combining YOLOv5s-DeepSORT, integrating target load recognition and trajectory tracking to improve adaptability to dock en...

Full description

Saved in:
Bibliographic Details
Main Authors: Chengzhi Wang, Donghong Chen, Zhen Liu, Yuanhao Li, Yifei Wang, Sanglan Zhao
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.0326376
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849715050705584128
author Chengzhi Wang
Donghong Chen
Zhen Liu
Yuanhao Li
Yifei Wang
Sanglan Zhao
author_facet Chengzhi Wang
Donghong Chen
Zhen Liu
Yuanhao Li
Yifei Wang
Sanglan Zhao
author_sort Chengzhi Wang
collection DOAJ
description To solve the problem of reduced positioning accuracy caused by changes in scale, background and occlusion in port and dock video images, this research proposes an enhanced model combining YOLOv5s-DeepSORT, integrating target load recognition and trajectory tracking to improve adaptability to dock environments. The findings indicate that incorporating multi-scale convolution into YOLOv5s improved the robustness of multi-scale object detection, resulting in a 0.4% increase in mean Average Precision (mAP). Furthermore, the integration of an efficient pyramid segmentation attention (EPSA) network enhanced the accuracy of multi-scale feature fusion representation. The model's mAP@0.5:0.95 increased by 1.2% following the introduction of EPSA. Finally, the original classification loss function was enhanced using a distributed sorting loss approach to mitigate the imbalance among loaded objects and the influence of background variations in the dock image sequence. This optimization led to a 3.1% improvement in multi-target tracking accuracy (MOTA). Experimental results on self-constructed datasets demonstrated an average accuracy of 90.9% and a detection accuracy of 92.2%, offering a valuable reference for target recognition and tracking in port and dock environments.
format Article
id doaj-art-1a1ea2b62f2f4364834d93da0d0648a9
institution DOAJ
issn 1932-6203
language English
publishDate 2025-01-01
publisher Public Library of Science (PLoS)
record_format Article
series PLoS ONE
spelling doaj-art-1a1ea2b62f2f4364834d93da0d0648a92025-08-20T03:13:31ZengPublic Library of Science (PLoS)PLoS ONE1932-62032025-01-01207e032637610.1371/journal.pone.0326376Port terminal mobile recognition based on combined YOLOv5s-DeepSort.Chengzhi WangDonghong ChenZhen LiuYuanhao LiYifei WangSanglan ZhaoTo solve the problem of reduced positioning accuracy caused by changes in scale, background and occlusion in port and dock video images, this research proposes an enhanced model combining YOLOv5s-DeepSORT, integrating target load recognition and trajectory tracking to improve adaptability to dock environments. The findings indicate that incorporating multi-scale convolution into YOLOv5s improved the robustness of multi-scale object detection, resulting in a 0.4% increase in mean Average Precision (mAP). Furthermore, the integration of an efficient pyramid segmentation attention (EPSA) network enhanced the accuracy of multi-scale feature fusion representation. The model's mAP@0.5:0.95 increased by 1.2% following the introduction of EPSA. Finally, the original classification loss function was enhanced using a distributed sorting loss approach to mitigate the imbalance among loaded objects and the influence of background variations in the dock image sequence. This optimization led to a 3.1% improvement in multi-target tracking accuracy (MOTA). Experimental results on self-constructed datasets demonstrated an average accuracy of 90.9% and a detection accuracy of 92.2%, offering a valuable reference for target recognition and tracking in port and dock environments.https://doi.org/10.1371/journal.pone.0326376
spellingShingle Chengzhi Wang
Donghong Chen
Zhen Liu
Yuanhao Li
Yifei Wang
Sanglan Zhao
Port terminal mobile recognition based on combined YOLOv5s-DeepSort.
PLoS ONE
title Port terminal mobile recognition based on combined YOLOv5s-DeepSort.
title_full Port terminal mobile recognition based on combined YOLOv5s-DeepSort.
title_fullStr Port terminal mobile recognition based on combined YOLOv5s-DeepSort.
title_full_unstemmed Port terminal mobile recognition based on combined YOLOv5s-DeepSort.
title_short Port terminal mobile recognition based on combined YOLOv5s-DeepSort.
title_sort port terminal mobile recognition based on combined yolov5s deepsort
url https://doi.org/10.1371/journal.pone.0326376
work_keys_str_mv AT chengzhiwang portterminalmobilerecognitionbasedoncombinedyolov5sdeepsort
AT donghongchen portterminalmobilerecognitionbasedoncombinedyolov5sdeepsort
AT zhenliu portterminalmobilerecognitionbasedoncombinedyolov5sdeepsort
AT yuanhaoli portterminalmobilerecognitionbasedoncombinedyolov5sdeepsort
AT yifeiwang portterminalmobilerecognitionbasedoncombinedyolov5sdeepsort
AT sanglanzhao portterminalmobilerecognitionbasedoncombinedyolov5sdeepsort