Multi-Stage Computer Vision Framework with Ensemble Learning for Real-Time Glass Packaging Defect Detection in Industrial Applications

Authors

  • Jonah Alfred Mekel Department of Mechanical Engineering, Mechatronics Program, Politeknik Negeri Manado, Manado, Indonesia
  • Rick Resa Wahani Department of Mechanical Engineering, Mechatronics Program, Politeknik Negeri Manado, Manado, Indonesia
  • Firmansyah Reskal Motulo Politeknik Negeri Manado
  • Alfred Noufie Mekel Department of Mechanical Engineering, Mechatronics Program, Politeknik Negeri Manado, Manado, Indonesia
  • Tineke Saroinsong Department of Mechanical Engineering, Production and Maintenance Program, Politeknik Negeri Manado, Manado, Indonesia
  • Tammy Tinny V. Pangow Department of Mechanical Engineering, Mechatronics Program, Politeknik Negeri Manado, Manado, Indonesia
  • Jerry Heisye Purnama Department of Mechanical Engineering, Mechatronics Program, Politeknik Negeri Manado, Manado, Indonesia
  • Jedithjah Naapia Tamedi Papia Department of Mechanical Engineering, Mechatronics Program, Politeknik Negeri Manado, Manado, Indonesia

DOI:

https://doi.org/10.59535/faase.v3i2.572

Keywords:

Computer Vision, Defect Detection, Glass Packaging, Ensemble Learning, Image Processing

Abstract

Transparent glass packaging inspection presents significant challenges for automated quality control systems due to optical complexities including reflections, refractions, and low-contrast defect patterns. This research develops a comprehensive multi-stage computer vision framework integrating specialized algorithmic modules with ensemble machine learning for real-time defect detection in industrial glass packaging lines. The framework implements four specialized detection stages: (1) meniscus-corrected liquid level measurement using dual-camera validation and polynomial surface fitting, (2) seal integrity assessment through Circular Hough Transform combined with geometric, texture, and color feature extraction, (3) lid positioning evaluation via calibrated geometric centroid analysis with tolerance-based classification, and (4) multi-method contamination detection integrating color aberration analysis, histogram-based particle detection, and morphological operations. The system employs an ensemble classification architecture combining modified MobileNetV2 convolutional neural network with Random Forest classifier, optimized for edge computing deployment. Industrial validation at PT AQUWAR Bintang Semesta demonstrated 91.6% overall detection accuracy with 347 milliseconds average processing time per container across 2,847 test samples spanning multiple defect categories. The modular framework architecture enables independent optimization of detection stages while maintaining real-time processing capabilities, providing a robust foundation for transparent packaging quality control in high-volume manufacturing environments.

Downloads

Download data is not yet available.

References

J. Smith, P. Brown, and C. Lee, “Challenges in transparent material quality control: A systematic review,” Int. J. Quality Control, vol. 29, no. 6, pp. 234–251, 2023.

H. Zhang, M. Liu, S. Wang, K. Chen, and J. Rodriguez, “High-speed vision inspection for beverage cup sealing quality,” IEEE Trans. Ind. Electron., vol. 71, no. 4, pp. 3251–3260, Apr. 2024.

V. Kumar, R. Singh, M. Patel, A. Thompson, and S. Lee, “Deep learning approaches for liquid level detection in transparent containers,” Pattern Recognit. Lett., vol. 168, pp. 45–52, 2024.

A. Pratama and S. Wijaya, “Computer vision implementation in Indonesian food and beverage industry: Challenges and opportunities,” J. Manuf. Technol. Manag., vol. 34, no. 2, pp. 189–207, 2023.

L. Chen and M. Wang, “Real-time quality control system for high-volume production lines using edge computing,” Comput. Ind., vol. 145, Art. no. 103812, 2023, doi: 10.1016/j.compind.2023.103812.

R. Johnson, D. Miller, and K. Brown, “Automated optical inspection systems for packaging applications: Design principles and performance metrics,” Opt. Eng., vol. 63, no. 2, Art. no. 024101, 2024.

C. Rodriguez, M. Garcia, P. Silva, and H. Tanaka, “Multi-modal sensor fusion for advanced quality control in manufacturing,” Sensors, vol. 23, no. 18, Art. no. 7892, 2023, doi: 10.3390/s23187892.

Cognex Corporation, “Machine vision applications in quality control: A comprehensive review,” Ind. Autom. Today, vol. 45, no. 3, pp. 112–128, 2023.

P. Anderson, J. Williams, and L. Davis, “Ensemble learning methods for industrial defect detection: A comparative study,” Mach. Learn. Manuf., vol. 8, no. 1, pp. 15–32, 2024.

J. Liu, A. Rodriguez, D. Kim, T. Nakamura, and F. Schmidt, “Ensemble methods for multi-defect detection in packaging industry,” Expert Syst. Appl., vol. 213, Art. no. 118932, 2024, doi: 10.1016/j.eswa.2024.118932.

M. Taylor and S. Jackson, “Transfer learning for industrial vision applications: Challenges and solutions,” Neural Netw., vol. 162, pp. 445–458, 2023.

X. Wang and A. Singh, “Adaptive learning approaches for manufacturing quality control systems,” IEEE Trans. Autom. Sci. Eng., vol. 21, no. 2, pp. 892–905, Apr. 2024.

K. Thompson, R. Clark, and E. Martinez, “Deep learning architectures for real-time industrial inspection,” Comput. Vis. Image Underst., vol. 238, Art. no. 103847, 2024.

A. G. Howard et al., “MobileNets: Efficient convolutional neural networks for mobile vision applications,” arXiv preprint, arXiv:1704.04861, 2017.

R. Thompson, M. Anderson, and D. Wilson, “Modified MobileNetV2 for edge computing in manufacturing applications,” Edge Comput. Res., vol. 5, no. 4, pp. 78–91, 2023.

J. Canny, “A computational approach to edge detection,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 8, no. 6, pp. 679–698, Nov. 1986, doi: 10.1109/TPAMI.1986.4767851.

R. O. Duda and P. E. Hart, “Use of the Hough transformation to detect lines and curves in pictures,” Commun. ACM, vol. 15, no. 1, pp. 11–15, Jan. 1972, doi: 10.1145/361237.361242.

R. M. Haralick, K. Shanmugam, and I. Dinstein, “Textural features for image classification,” IEEE Trans. Syst., Man, Cybern., vol. 3, no. 6, pp. 610–621, Nov. 1973, doi: 10.1109/TSMC.1973.4309314.

N. Otsu, “A threshold selection method from gray-level histograms,” IEEE Trans. Syst., Man, Cybern., vol. 9, no. 1, pp. 62–66, Jan. 1979, doi: 10.1109/TSMC.1979.4310076.

L. Breiman, “Random forests,” Mach. Learn., vol. 45, no. 1, pp. 5–32, 2001, doi: 10.1023/A:1010933404324.

Downloads

Published

2025-12-31

How to Cite

Jonah Alfred Mekel, Rick Resa Wahani, Motulo, F. R., Alfred Noufie Mekel, Tineke Saroinsong, Tammy Tinny V. Pangow, Jerry Heisye Purnama, & Jedithjah Naapia Tamedi Papia. (2025). Multi-Stage Computer Vision Framework with Ensemble Learning for Real-Time Glass Packaging Defect Detection in Industrial Applications. Frontier Advances in Applied Science and Engineering, 3(2), 81–98. https://doi.org/10.59535/faase.v3i2.572

Issue

Vol. 3 No. 2 (2025)

Section

Original Articles

License

Copyright (c) 2026 Jonah Alfred Mekel, Rick Resa Wahani, Firmansyah Reskal Motulo, Alfred Noufie Mekel, Tineke Saroinsong, Tammy Tinny V. Pangow, Jerry Heisye Purnama, Jedithjah Naapia Tamedi Papia

Creative Commons License

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

By submitting a manuscript to Frontier Advances in Applied Science and Engineering (FAASE), the author(s) agree to the following terms:

  1. Copyright: The author(s) retain copyright to their work, granting Frontier Advances in Applied Science and Engineering (FAASE) the right to publish it. Articles are licensed under a Creative Commons Attribution 4.0 International License (CC BY).

  2. License Grant: The CC BY license allows others to share and adapt the work for any purpose, even commercially, provided proper attribution is given to the original author(s) and the source, a link to the Creative Commons license is provided, and any changes made are indicated.

  3. Author's Rights: Authors are allowed to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work, with proper acknowledgment of its initial publication in Frontier Advances in Applied Science and Engineering (FAASE).

  4. Permissions: Reproduction, posting, transmission, or other distribution or use of the final published article, in whole or in part, requires prior written permission from Frontier Advances in Applied Science and Engineering (FAASE) and the respective author(s).

  5. Attribution: Proper attribution must be given to the original author(s) and the source, indicating that the work was originally published in Frontier Advances in Applied Science and Engineering (FAASE) with a link to the published article.

  6. No Warranty: The journal and its editorial team make no representations or warranties regarding the accuracy, completeness, or fitness for a particular purpose of the published work.

  7. Dispute Resolution: Any disputes arising from the use of the licensed material shall be governed by the laws of [Jurisdiction] without regard to conflicts of law principles.

By submitting a manuscript to FAASE, the author(s) agree to abide by these license terms and grant the necessary rights to Frontier Advances in Applied Science and Engineering (FAASE) for the publication and dissemination of the submitted work.