Skip to main content Skip to journal navigation menu Skip to site footer
contact@ukscip.com
Article

Research on Industrial Defect Machine Vision Detection Algorithm Based on Deep Learning and Spectral Fusion

Kaisheng Deng ORCID
    Received: 5 February 2026; Revised: 27 March 2026; Accepted: 6 April 2026; Published: 31 August 2026

    Abstract

    Industrial surface defect detection is core to intelligent manufacturing quality assurance, guaranteeing stable product quality and reliable production. Conventional defect detection deep learning algorithms relying on Red-Green-Blue (RGB) images suffer low accuracy and weak robustness amid complicated industrial scenes. This paper puts forward Multi-Spectral Fusion Defect Detection Network (MS-FDNet), a novel industrial surface defect detection framework combining deep learning and spectral imaging. It adopts a spectral-spatial attention coupling module to fully fuse multi-band spectral information and exploit distinctive feature advantages of each band. Self-calibrated convolution coupled with spatial pyramid structure constructs a multi-scale feature extraction branch to strengthen the model’s perception of defects of varying sizes. A multi-task composite loss integrating Focal Loss and Dice Loss is designed to balance model memory occupation and detection precision and boost overall inference efficiency. Tests on the public NEU-det dataset reveal the proposed algorithm achieves a mean Average Precision at Intersection over Union threshold 0.5 (mAP@0.5) of 81.1%, a 4.6% improvement over the baseline while retaining fast inference. It runs at 67.9 frames per second with a tiny model volume of merely 4.22 MB, showing outstanding deployment practicability. The method greatly enhances defect identification performance and system stability under uneven illumination and cluttered backgrounds. This work offers an innovative industrial defect detection scheme with valuable theoretical research value and promising engineering application potential.

    Keywords

    References

    1. 1. López de la Rosa, F.; Gómez-Sirvent, J.L.; Morales, R.; et al. Defect detection and classification on semiconductor wafers using two-stage geometric transformation-based data augmentation and SqueezeNet lightweight convolutional neural network. Comput. Ind. Eng. 2023, 183, 109549.
    2. 2. Khanam, R.; Hussain, M.; Hill, R. PDNet: A lightweight attention-guided CNN for efficient pallet racking defect detection on edge devices. Discover Artif. Intell. 2025, 5, 309.
    3. 3. Duarte, J.; Claro, F.M.; Vitoriano, A.M.P.; et al. Detection and Classification of Defects on Metal Surfaces Based on a Lightweight YOLOX-Tiny COCO Network. Eng. 2025, 6, 302.
    4. 4. Rodríguez, E.; Pauccar, G.; Carrizales, J.; et al. Lightweight deep learning framework for quality assessment of rolled fondant. LWT 2026, 252, 119580.
    5. 5. Raj, A.A.V.; Pathmanaban, P.; Raja, G.M.; et al. Lightweight deep learning models for visual weld defect inspection: Balancing speed and accuracy with YOLO architectures. Weld. Int. 2026, 40, 70–84.
    6. 6. Vijayakumar, A.; Koilraj, S.A.J.; Rajappa, M.; et al. LightDefectNet-18: A Lightweight Framework for Multi-Domain Defect Detection. J. Nondestruct. Eval. 2026, 45, 27.
    7. 7. Ding, Z.; Yue, X.; Zeng, F.; et al. Spectral Detection of Maize Seed Vitality Based on Machine Learning and Deep Learning. J. Huazhong Agric. Univ. 2023, 42, 230–240. (in Chinese)
    8. 8. Gao, H.; Guo, Z.; Zeng, Y.; et al. Early classification and detection of soft rot disease in kiwifruit based on hyperspectral image band fusion. Spectrosc. Spectr. Anal. 2024, 44, 241–249. (in Chinese)
    9. 9. Deng, Y.; Huang, Z. Multispectral image fusion method for surface defect detection of IC devices. Opt. Precis. Eng. 2024, 32, 740–751. (in Chinese)
    10. 10. Chen, J.; Li, C.; Huang, H.; et al. Computer motherboard assembly defect detection using parallel feature extraction and progressive feature fusion. Opt. Precis. Eng. 2024, 32, 1622–1637. (in Chinese)
    11. 11. Moshayedi, A.J.; Khan, A.S.; Zhang, M.J.; et al. MobileNetV3 for Pipeline Welding Defect Detection: A Computer Vision Approach on Noisy RGB Images. E3S Web Conf. 2026, 723, 01014.
    12. 12. Czimmermann, T.; Ciuti, G.; Milazzo, M.; et al. Visual-Based Defect Detection and Classification Approaches for Industrial Applications—A Survey. Sensors 2020, 20, 1459.
    13. 13. Saberironaghi, A.; Ren, J.; El-Gindy, M. Defect Detection Methods for Industrial Products Using Deep Learning Techniques: A Review. Algorithms 2023, 16, 95.
    14. 14. Shin, H.; Ahn, Y.; Song, M.; et al. Visualization for Explanation of Deep Learning-Based Defect Detection Model Using Class Activation Map. Comput. Mater. Contin. 2023, 75, 4753–4766.
    15. 15. Maram, B.; Kumar, S.K.; Gampala, V. Multispectral and hyperspectral image fusion: A systematic analysis and review with the state of art techniques. Int. J. Bio-Inspired Comput. 2024, 23, 214–224.
    16. 16. Fisne, A.; Kalay, A.; Eken, S. Fast and efficient computing for deep learning-based defect detection models in lightweight devices. J. Intell. Manuf. 2024, 36, 5035–5050.
    17. 17. Kılıç, K.; Kılıç, K.; Doğru, İ.A.; et al. WD Detector: Deep learning-based hybrid sensor design for wood defect detection. Eur. J. Wood Wood Prod. 2025, 83, 50.
    18. 18. Kim, J.; Kim, J.; Lian, Y.; et al. Point Cloud-Based Defect Inspection with Multisensor Fusion and Deep Learning for Advancing Building Construction Quality. J. Civil Eng. Comput. 2025, 39, 93–107.
    19. 19. Hassan, A.M.; Hassan, M.; Sadek, A.; et al. Focus to generalize (F2G): Physics-guided attention for sample efficient and generalizable deep learning defect detection. J. Intell. Manuf. 2025, 37, 1789–1812.
    20. 20. Atiya, N.; Shemer, A.; Schwarz, A.; et al. Imaging through Scattering Tissue Based on NIR Multispectral Image Fusion Technique. Sensors 2025, 25, 4977.
    21. 21. Lee, M.; Choi, J. ID-RAG: Industrial defect retrieval-augmented generation for industrial surface defect detection. Mach. Vis. Appl. 2026, 37, 36.
    22. 22. Schmedemann, O.; Baaß, M.; Schoepflin, D.; et al. Procedural synthetic training data generation for AI-based defect detection in industrial surface inspection. Procedia CIRP 2022, 107, 1101–1106.
    23. 23. Siami, M.; Dąbek, P.; Shiri, H.; et al. A Surface Defect Detection System for Industrial Conveyor Belt Inspection Using Apple’s TrueDepth Camera Technology. Appl. Sci. 2026, 16, 609.
    24. 24. Shin, J.; Hamza, M.; Kim, H.; et al. Lightweight Swin Transformer for high-precision industrial defect detection in smart manufacturing. Alexandria Eng. J. 2025, 130, 227–240.
    25. 25. Ancha, K.V.; Gonuguntla, V.; Vaddi, R. TRSBi-YOLO: Transformer based lightweight and high-performance model for PCB defects detection. J. Supercomput. 2025, 81, 1277.
    ×