Laser weld spot detection based on YOLO-weld

Abstract Laser weld point detection is crucial in modern industrial manufacturing, yet it faces challenges such as a limited number of samples, uneven distribution, and diverse, irregular shapes. To address these issues, this paper proposes an innovative model, YOLO-Weld, which achieves lightweight...

Full description

Saved in:
Bibliographic Details
Main Authors: Jianxin Feng, Jiahao Wang, Xinyu Zhao, Zhiguo Liu, Yuanming Ding
Format: Article
Language:English
Published: Nature Portfolio 2024-11-01
Series:Scientific Reports
Subjects:
Online Access:https://doi.org/10.1038/s41598-024-80957-3
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850064659851247616
author Jianxin Feng
Jiahao Wang
Xinyu Zhao
Zhiguo Liu
Yuanming Ding
author_facet Jianxin Feng
Jiahao Wang
Xinyu Zhao
Zhiguo Liu
Yuanming Ding
author_sort Jianxin Feng
collection DOAJ
description Abstract Laser weld point detection is crucial in modern industrial manufacturing, yet it faces challenges such as a limited number of samples, uneven distribution, and diverse, irregular shapes. To address these issues, this paper proposes an innovative model, YOLO-Weld, which achieves lightweight design while enhancing detection accuracy. Firstly, a targeted data augmentation strategy is employed to increase both the quantity and diversity of samples from minority classes. Following this, a Diverse Class Normalization Loss (DCNLoss)function is designed to emphasize the importance of tail data in the model’s training. Secondly, the Adaptive Hierarchical Intersection over Union Loss (AHIoU Loss)function is introduced, which assigns varying levels of attention to different Intersections over Union (IoU) samples, with a particular focus on moderate IoU samples, thereby accelerating the bounding box regression process. Finally, a lightweight multi-scale feature processing module, MSBCSPELAN, is proposed to enhance multi-scale feature handling while reducing the number of model parameters. Experimental results indicate that YOLO-Weld significantly improves the accuracy and efficiency of laser weld point detection, with mean Average Precision at 50 ( $$\:{mAP}_{50}$$ ) and mean Average Precision at 50:95 ( $$\:{mAP}_{50:95}$$ ) increasing by 15.6% and 15.8%, respectively. Additionally, the model’s parameter count is reduced by 0.4 M, GFLOPS decreases by 1.1, precision improves by 4.3%, recall rises by 22.2%, and the F1 score increases by 15.1%.
format Article
id doaj-art-6fca707167744062b45cada5eabc6abd
institution DOAJ
issn 2045-2322
language English
publishDate 2024-11-01
publisher Nature Portfolio
record_format Article
series Scientific Reports
spelling doaj-art-6fca707167744062b45cada5eabc6abd2025-08-20T02:49:15ZengNature PortfolioScientific Reports2045-23222024-11-0114111810.1038/s41598-024-80957-3Laser weld spot detection based on YOLO-weldJianxin Feng0Jiahao Wang1Xinyu Zhao2Zhiguo Liu3Yuanming Ding4Communication and Network Laboratory, Dalian UniversityCommunication and Network Laboratory, Dalian UniversityCommunication and Network Laboratory, Dalian UniversityCommunication and Network Laboratory, Dalian UniversityCommunication and Network Laboratory, Dalian UniversityAbstract Laser weld point detection is crucial in modern industrial manufacturing, yet it faces challenges such as a limited number of samples, uneven distribution, and diverse, irregular shapes. To address these issues, this paper proposes an innovative model, YOLO-Weld, which achieves lightweight design while enhancing detection accuracy. Firstly, a targeted data augmentation strategy is employed to increase both the quantity and diversity of samples from minority classes. Following this, a Diverse Class Normalization Loss (DCNLoss)function is designed to emphasize the importance of tail data in the model’s training. Secondly, the Adaptive Hierarchical Intersection over Union Loss (AHIoU Loss)function is introduced, which assigns varying levels of attention to different Intersections over Union (IoU) samples, with a particular focus on moderate IoU samples, thereby accelerating the bounding box regression process. Finally, a lightweight multi-scale feature processing module, MSBCSPELAN, is proposed to enhance multi-scale feature handling while reducing the number of model parameters. Experimental results indicate that YOLO-Weld significantly improves the accuracy and efficiency of laser weld point detection, with mean Average Precision at 50 ( $$\:{mAP}_{50}$$ ) and mean Average Precision at 50:95 ( $$\:{mAP}_{50:95}$$ ) increasing by 15.6% and 15.8%, respectively. Additionally, the model’s parameter count is reduced by 0.4 M, GFLOPS decreases by 1.1, precision improves by 4.3%, recall rises by 22.2%, and the F1 score increases by 15.1%.https://doi.org/10.1038/s41598-024-80957-3Laser Weld Spot DetectionYOLO-WeldThe long tail effectBounding box regressionMulti-scale features
spellingShingle Jianxin Feng
Jiahao Wang
Xinyu Zhao
Zhiguo Liu
Yuanming Ding
Laser weld spot detection based on YOLO-weld
Scientific Reports
Laser Weld Spot Detection
YOLO-Weld
The long tail effect
Bounding box regression
Multi-scale features
title Laser weld spot detection based on YOLO-weld
title_full Laser weld spot detection based on YOLO-weld
title_fullStr Laser weld spot detection based on YOLO-weld
title_full_unstemmed Laser weld spot detection based on YOLO-weld
title_short Laser weld spot detection based on YOLO-weld
title_sort laser weld spot detection based on yolo weld
topic Laser Weld Spot Detection
YOLO-Weld
The long tail effect
Bounding box regression
Multi-scale features
url https://doi.org/10.1038/s41598-024-80957-3
work_keys_str_mv AT jianxinfeng laserweldspotdetectionbasedonyoloweld
AT jiahaowang laserweldspotdetectionbasedonyoloweld
AT xinyuzhao laserweldspotdetectionbasedonyoloweld
AT zhiguoliu laserweldspotdetectionbasedonyoloweld
AT yuanmingding laserweldspotdetectionbasedonyoloweld