Education, Science, Technology, Innovation and Life
Open Access
Sign In

Algorithm for Surface Defect Detection of Aluminum Profiles Based on Improved YOLOv11

Download as PDF

DOI: 10.23977/autml.2025.060117 | Downloads: 11 | Views: 521

Author(s)

Xinrui Chen 1, Zijian Dong 1, Jiangzheng Xu 1, Zhaojin Huangfu 1

Affiliation(s)

1 School of Electronic Engineering, Jiangsu Ocean University, Lianyungang, Jiangsu, 222000, China

Corresponding Author

Zijian Dong

ABSTRACT

The surface defect detection of aluminum profiles is a crucial core link in industrial quality control. Traditional detection methods have practical problems such as easy missed detection of small-target defects and easy misjudgment in complex texture scenarios. This study constructs an efficient detection model suitable for the surface of aluminum profiles based on the YOLOv11 algorithm framework. EfficientViT is used to reconstruct the backbone network, and the multi-scale feature extraction capability is enhanced through hierarchical attention mechanisms and lightweight convolution operations. A Dynamic Deformable Feature Pyramid Network (DDFPN) is introduced, integrating RepGFPN re-parameterized connections, VoVGSCSP grouped convolution, and CoordAtt coordinate attention mechanism to achieve adaptive fusion of defect features and directional sensitivity perception. Experimental results show that compared with the original model, the improved EDHF-YOLO model significantly improves detection accuracy and greatly reduces calculation amount, effectively balancing detection performance and computational efficiency, and providing an innovative technical solution for surface defect detection of aluminum profiles.

KEYWORDS

Defect Detection, EDHF-YOLO, RepGFPN, Dynamic Deformable Feature Pyramid

CITE THIS PAPER

Xinrui Chen, Zijian Dong, Jiangzheng Xu, Zhaojin Huangfu, Algorithm for Surface Defect Detection of Aluminum Profiles Based on Improved YOLOv11. Automation and Machine Learning (2025) Vol. 6: 146-155. DOI: http://dx.doi.org/10.23977/autml.2025.060117.

REFERENCES

[1] LU J Z, ZHANG Y C, LIU S P, et al. Lightweight DCN-YOLO for Strip Surface Defect Detection in Complex Environments [J]. Computer Engineering and Applications,2023, 59(15):318-328.
[2] WU L, CHU Y K, YANG H G, et al. Sim-YOLOv8 Object Detection Model for DR Image Defects in Aluminum Alloy Welds[J]. Chinese Journal of Lasers,2024,51(16): 29-38.
[3] Li B L. Design and research of automatic cloth sewing machine[D]. Shanghai: Donghua University, 2022.
[4] WU Z H, ZHONG M E, TAN J W, et al. Research on five types of typical defects image detection algorithms for complex textured fabrics[J]. Electronic Measurement Technology, 2023, 46(16): 57-63.
[5] XU Y D, CAI Y H, LI Y, et al. Lightweight overhead transmission line bird′s nest detection network based on YOLOv5s[J]. Electronic Measurement Technology, 2024, 47(7): 138-148.
[6] DONG C, ZHANG K, XIE Z Y, et al. An improved cascade RCNN detection method for key components and defects of transmission lines[J]. IET Generation, Transmission Distribution, 2023, 17(19): 427-439.
[7] ZENG Q, LI B. Cucumber detection algorithm based on improved SSD[J]. Foreign Electronic Measurement Technology, 2023, 42(04): 158-165.
[8] ZHENG X , SHAO Z ,CHEN Y , et al. MSPB-YOLO: High-Precision Detection Algorithm of Multi-Site Pepper Blight Disease Based on Improved YOLOv8[J].Agronomy,2025,15(4):839-858.
[9] DONG C ,SHEN Y ,FENG Z , et al. Connecting finger defects in flexible touch screen inspected with machine vision based on YOLOv8n[J].Measurement,2025, 17(19):246-254.
[10] TU X K, ZHENG S W, YU S H, el at. 3D object detection network based on symmetric shape generation[J]. Chinese Journal of Scientific Instrument, 2023, 44(6): 252-263.
[11] XU F X, FAN R, MA X L. Improved YOLOv7 algorithm for crowded pedestrian detection[J]. Computer Engineering, 2024, 50(3): 250-258.

Downloads: 3870
Visits: 177442

Sponsors, Associates, and Links


All published work is licensed under a Creative Commons Attribution 4.0 International License.

Copyright © 2016 - 2031 Clausius Scientific Press Inc. All Rights Reserved.