A 3D printing defect detection algorithm incorporating attention mechanism

ZHANG Junjie; SHEN Zhen; FANG Qihang; DONG Xisong; WANG Di; XIONG Gang

doi:10.11959/j.issn.2096-3750.2025.00464

您当前的位置：

首页 >

文章列表页 >

A 3D printing defect detection algorithm incorporating attention mechanism

Theory and Technology | 更新时间：2026-01-26

- A 3D printing defect detection algorithm incorporating attention mechanism
- Chinese Journal on Internet of Things Vol. 9, Issue 3, Pages: 132-142(2025)
- 作者机构：
  
  1.中国科学院自动化研究所多模态人工智能系统全国重点实验室，北京 100190
  2.中国科学院大学人工智能学院，北京 100049
  3.中国科学院自动化研究所北京市智能化技术与系统工程技术研究中心，北京 100190
  4.华南理工大学机械与汽车工程学院，广东广州 510641
- 作者简介：
- 基金信息：
  
  The National Natural Science Foundation of China(92267103;92360307;62461160259)
- DOI：10.11959/j.issn.2096-3750.2025.00464
  CLC： TP391.41
- Received：15 November 2024，
  
  Revised：2025-01-08，
  
  Published：10 September 2025
- 稿件说明：
移动端阅览
张俊杰,沈震,方启航等.融合注意力机制的3D打印缺陷检测算法[J].物联网学报,2025,09(03):132-142.

ZHANG Junjie,SHEN Zhen,FANG Qihang,et al.A 3D printing defect detection algorithm incorporating attention mechanism[J].Chinese Journal on Internet of Things,2025,09(03):132-142.
张俊杰,沈震,方启航等.融合注意力机制的3D打印缺陷检测算法[J].物联网学报,2025,09(03):132-142. DOI： 10.11959/j.issn.2096-3750.2025.00464.

ZHANG Junjie,SHEN Zhen,FANG Qihang,et al.A 3D printing defect detection algorithm incorporating attention mechanism[J].Chinese Journal on Internet of Things,2025,09(03):132-142. DOI： 10.11959/j.issn.2096-3750.2025.00464.

摘要

近年来，3D打印技术在越来越多的行业中发挥着重要作用，但是，其作为一种新技术，相对于传统制造，在打印过程中出现缺陷的情形较多，这些缺陷会显著影响最终产品的性能。鉴于3D打印部件通常具有复杂和高度优化的几何形状，传统的检测技术难以满足其对精确性和效率的要求。为解决这一难题，提出了一种基于改进版YOLOv5的3D打印缺陷检测算法。该算法对YOLOv5模型进行了深入改进，通过替换损失函数和引入注意力机制，成功实现了模型的轻量化。新设计的检测系统具有参数规模较小、推理速度快、检测精度高和鲁棒性较好的特点。与原始的YOLOv5s模型相比，改进后的轻量化模型在检测精度上达到了94.2%，同时在参数规模和数据大小上几乎减半。这一进步不仅提升了检测效率，还为3D打印缺陷检测与故障诊断提供了一种有效的技术手段。

Abstract

In recent years

3D printing technology has played an increasingly important role in a growing number of industries. However

as a relatively new technology

it tends to exhibit more defects during the printing process compared to traditional manufacturing methods. These defects can significantly impact the performance of the final product. Given that 3D printed parts typically have complex and highly optimized geometric shapes

traditional detection technologies struggle to meet the demands for precision and efficiency. To address this challenge

this paper introduces a 3D printing defect detection algorithm based on an improved version of YOLOv5. The algorithm makes extensive refinements to the YOLOv5 model

achieving model lightweighting by replacing the loss function and introducing an attention mechanism. The newly designed detection system is characterized by a smaller parameter scale

rapid inference speed

high detection accuracy

and strong robustness. Compared to the original YOLOv5s model

the improved lightweight model has achieved a detection accuracy of 94.2%

and has nearly halved the parameter scale. This advancement not only enhances detection efficiency but also provides an effective technical solution for 3D printing defect detection and fault diagnosis.

关键词

Keywords

references

喻朝新 , 郭松涛 , 郭佳哲 . 基于工业互联技术的胶料质量缺陷检测模型构建分析 [J ] . 粘接 , 2023 , 50 ( 6 ): 19 - 22 .

YU C X , GUO S T , GUO J Z . Research on rubber defect identification and detection based on industrial interconnection technology [J ] . Adhesion , 2023 , 50 ( 6 ): 19 - 22 .

袁朕鑫 . 工业互联网中边缘端轮辋焊缝检测系统设计与实现 [D ] . 济南：济南大学 , 2023 .

YUAN Z X . Design and implementation of edge end rim weld detection system in industrial Internet [D ] . Jinan : University of Jinan , 2023 .

袁慧苗 . 基于工业互联网的旋转机械系统状态检测与故障诊断研究 [D ] . 济南：齐鲁工业大学 , 2023 .

YUAN H M . Research on condition detection and fault diagnosis of rotating machinery system based on industrial Internet . [D ] . Jinan : Qilu University of Technology , 2023 .

佟星 . 工业互联网场景下基于人工智能的图像和数据分析研究 [D ] . 广州：广东技术师范大学 , 2021 .

TONG X . Image and data analysis based on artificial intelligencein industrial Internet . [D ] . Guangzhou : Guangdong University of Technology and Education , 2021 .

陈立君 . 基于流形支持向量机的木材表面缺陷识别方法的研究 [D ] . 哈尔滨：东北林业大学 , 2015 .

CHEN L J . The research for recognition method of wood surface defectbased on svm combined manifold . [D ] . Harbin : Northeast Forestry University , 2015 .

刘浩 , 陈再良 , 王善翔 . 基于改进YOLOv3算法的刀具表面缺陷检测 [J ] . 组合机床与自动化加工技术 , 2021 ( 11 ): 87 - 90 .

LIU H , CHEN Z L , WANG S X . Tool surface defect detection based on improved YOLOv3 algorithm [J ] . Modular Machine Tool & Automatic Manufacturing Technique , 2021 ( 11 ): 87 - 90 .

QIU J H , HU Y H , CUI J R , et al . Textile defect classification based on convolutional neural network and SVM [J ] . AATCC Journal of Research , 2021 , 8 ( 1_suppl ): 75 - 81 .

舒琪 . 基于机器视觉的铆钉筛选系统设计 [D ] . 成都：西南交通大学 , 2022 .

SHU Q . Design of a rivet screening system based on machine vision [D ] . Chengdu : Southwest Jiaotong University , 2022 .

YU Z L , LEI Y Q , SHEN F , et al . Application of improved YOLOv5 algorithm in lightweight transmission line small target defect detection [J ] . Electronics , 2024 , 13 ( 2 ):

HU B , WANG J H . Detection of PCB surface defects with improved faster-RCNN and feature pyramid network [J ] . IEEE Access , 2020 , 8 : 108335 - 108345 .

HU Y M , WEN B , YE Y S , et al . Multi-defect detection network for high-voltage insulators based on adaptive multi-attention fusion [J ] . Applied Sciences , 2023 , 13 ( 24 ): 13351 .

YU J , JIANG Y , WANG Z , et al . UnitBox: an advanced object detection network [J ] . arXiv preprint , 2016 , arXiv: 1608.01471 .

赵英伟 . 面向工业场景的机器学习算法应用研究 [D ] . 成都：电子科技大学 , 2023 .

ZHAO Y W . Machine learning algorithms for industrial scenarios applied research . [D ] . Chengdu : University of Electronic Science and Technology of China , 2023 .

REZATOFIGHI H , TSOI N , GWAK J , et al . Generalized intersection over union: a metric and a loss for bounding box regression [C ] // Proceedings of the 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) . Piscataway : IEEE Press , 2019 : 658 - 666 .

ZHENG Z H , WANG P , LIU W , et al . Distance-IoU loss: faster and better learning for bounding box regression [J ] . Proceedings of the AAAI Conference on Artificial Intelligence , 2020 , 34 ( 7 ): 12993 - 13000 .

ZHANG Y F , REN W Q , ZHANG Z , et al . Focal and efficient IOU loss for accurate bounding box regression [J ] . Neurocomputing , 2022 , 506 : 146 - 157 .

GEVORGYAN Z . SIoU loss: more powerful learning for bounding box regression [J ] . arXiv preprint , 2022 ,arXiv: 2205.12740 .

YANG L X , ZHANG R Y , LI L D , et al . SimAM: a simple, parameter-free attention module for convolutional neural networks [C ] // Proceedings of the International Conference on Machine Learning , 2021 .

HU J , SHEN L , SUN G . Squeeze-and-excitation networks [C ] // Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition . Piscataway : IEEE Press , 2018 : 7132 - 7141 .

ZHANG Q L , YANG Y B . SA-net: shuffle attention for deep convolutional neural networks [C ] // Proceedings of the ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) . Piscataway : IEEE Press , 2021 : 2235 - 2239 .

HAN K , WANG Y H , TIAN Q , et al . GhostNet: more features from cheap operations [C ] // Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) . Piscataway : IEEE Press , 2020 : 1577 - 1586 .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Feature fusion method for UWB angle of arrival estimation under non-line-of-sight conditions

Lightweight parallel denoising network-based OTFS channel estimation algorithm

Review of the application of deep learning in the auxiliary diagnosis of ophthalmic diseases

Model pruning techniques in the Internet of things: state of the art, methods and perspectives

Research on EEG signal classification of motor imagery based on AE and Transformer

Related Author

LI Jianjia

ZHANG Tingting

YIN Jikai

BAO Yachuan

LIU Jinting

ZHENG Bofeng

ZENG Wenmin

ZHANG Hui

Related Institution

Guangdong Provincial Key Laboratory of Space-Aerial Networking and Intelligent Sensing, Harbin Institute of Technology

The 54th Research Institute of China Electronics Technology Group Corporation

State Key Laboratory of Comprehensive PNT Network and Equipment Technology, The 54th Research Institute of China Electronics Technology Group Corporation

IOT Thrust, The Hong Kong University of Science and Technology

College of Electronic Science and Technology, National University of Defense Technology

⁰