Performance Evaluation of Machine Vision-Based Pouch Packaging Defect Detection System on MSME Production Line
Article Sidebar
Main Article Content
Abstract
Manual packaging quality inspection has limitations regarding subjectivity and consistency, especially for high production volumes. This study developed an automated defect-detection system based on machine vision to distinguish between normal and defective pouch packaging. The proposed approach extracted shape and texture features from packaging images to train three machine learning algorithms: Random Forest, Support Vector Machine (SVM), and logistic regression. Model evaluation showed that the Random Forest algorithm achieved the highest accuracy rate of 100%, followed by logistic regression at 99%, and SVM at 97%. Further analysis revealed that aspect ratio, with a 35% contribution, and circularity at 19% were the primary determinants in the classification process. Although the accuracy rate was exceptionally high, diagnostic checks revealed evidence of data leakage due to inconsistent image acquisition conditions. The conclusion of this study indicates that the system has substantial potential for industrial automation, yet operational implementation requires strict standardization of image acquisition to prevent the models from learning shortcuts.
Downloads
Article Details
Section
This work is licensed under a Creative Commons Attribution 4.0 International License.
References
D’Angelo, G., & Rampone, S. (2015). Shape-based defect classification for non destructive testing. 2nd IEEE International Workshop on Metrology for Aerospace, MetroAeroSpace 2015 - Proceedings, 406–410. https://doi.org/10.1109/METROAEROSPACE.2015.7180691
Fu, Y., Downey, A. R. J., Yuan, L., Zhang, T., Pratt, A., & Balogun, Y. (2022). Machine learning algorithms for defect detection in metal laser-based additive manufacturing: A review. Journal of Manufacturing Processes, 75, 693–710. https://doi.org/10.1016/J.JMAPRO.2021.12.061
Geirhos, R., Jacobsen, J. H., Michaelis, C., Zemel, R., Brendel, W., Bethge, M., & Wichmann, F. A. (2020). Shortcut learning in deep neural networks. Nature Machine Intelligence 2020 2:11, 2(11), 665–673. https://doi.org/10.1038/s42256-020-00257-z
Hilbe, J. M. (2025). Logistic Regression. International Encyclopedia of Statistical Science, 1386–1390. https://doi.org/10.1007/978-3-662-69359-9_333
Kim, D., Kim, H., Kim, Y., Chae, M., Ko, Y. M., Bae, Y. M., Sim, H., Oh, Y. C., & Noh, K. H. (2025). Using statistical models for optimal packaging in semiconductor manufacturing processes. Journal of the Korean Statistical Society, 54(1), 1–19. https://doi.org/10.1007/S42952-024-00284-1/TABLES/5
Kim, T., & Behdinan, K. (2022). Advances in machine learning and deep learning applications towards wafer map defect recognition and classification: a review. Journal of Intelligent Manufacturing 2022 34:8, 34(8), 3215–3247. https://doi.org/10.1007/S10845-022-01994-1
Lee, Y. J., Joo, M. J., Yu, H. K., Lee, T. J., & Kim, H. J. (2025). Random forest regressor for predicting sensory texture of emotional designed packaging films. Results in Engineering, 25, 104147. https://doi.org/10.1016/J.RINENG.2025.104147
Liu, B., Chen, Y., Xie, J., & Chen, B. (2021). Industrial Printing Image Defect Detection Using Multi-Edge Feature Fusion Algorithm. Complexity, 2021(1), 2036466. https://doi.org/10.1155/2021/2036466
Liu, G., Zhang, S., Wang, L., Li, X., & Li, G. (2025). Research on mechanical automatic food packaging defect detection model based on improved YOLOv5 algorithm. PLOS ONE, 20(4), e0321971. https://doi.org/10.1371/JOURNAL.PONE.0321971
Minghui, G., Jun, Z., & Huijuan, L. (2023). Research progress of food packaging defect detection based on machine vision. Food & Machinery, 39(9), 95-102,116. https://doi.org/10.13652/J.SPJX.1003.5788.2022.80943
Pauline, N., Tamala, I., Ariella Butar Butar, C., Jaya, A. M., Nathania, E. G., Made, N., Kireina Maharani, P., Sihotang, C. H., Hamdy, D., & Destyanto, Y. R. T. (2026). Model Prediksi Cacat Produk sebagai Early Detection System pada Quality Control Manufaktur. In D. B. Setyohadi (Ed.), Prosiding Seminar Nasional KONSTELASI (Vol. 3, Number 1, pp. 224–235). Universitas Atma Jaya Yogyakarta. https://doi.org/10.24002/PROSIDINGKONSTELASI.V3I1.13539
Son, G. J., Jung, H. C., & Kim, Y. D. (2024). Temporal-Quality Ensemble Technique for Handling Image Blur in Packaging Defect Inspection. Sensors 2024, Vol. 24, Page 4438, 24(14), 4438. https://doi.org/10.3390/S24144438
Susanti, N., Widodo, C. E., & Setiawan, A. (2025). A Systematic Review of AI-based Packaging Defect Detection to Strengthen Digital Industrial Resilience and Sustainability. Proceedings - 7th International Conference on Informatics, Multimedia, Cyber and Information System, ICIMCIS 2025, 378–383. https://doi.org/10.1109/ICIMCIS68501.2025.11326917
Tabernik, D., Šela, S., Skvarč, J., & Skočaj, D. (2019). Segmentation-based deep-learning approach for surface-defect detection. Journal of Intelligent Manufacturing 2019 31:3, 31(3), 759–776. https://doi.org/10.1007/S10845-019-01476-X
Valkenborg, D., Rousseau, A. J., Geubbelmans, M., & Burzykowski, T. (2023). Support vector machines. American Journal of Orthodontics and Dentofacial Orthopedics, 164(5), 754–757. https://doi.org/10.1016/j.ajodo.2023.08.003
Vanany, I., Hua Tan, K., Siswanto, N., Arvitrida, N. I., & Pahlawan, F. M. (2021). Halal six sigma framework for defects reduction. Journal of Islamic Marketing, 12(4), 776–793. https://doi.org/10.1108/JIMA-11-2019-0232
Wu, Y., & Lu, Y. (2019). An intelligent machine vision system for detecting surface defects on packing boxes based on support vector machine. Measurement and Control (United Kingdom), 52(7–8), 1102–1110. https://doi.org/10.1177/0020294019858175;ISSUE:ISSUE:DOI
Yahyaei, K., Khan, M. S. M., Ghosh, S., Roy, A., Craig, P., Varshney, N., & Asadizanjani, N. (2025). X-ADAPT: AI-driven design-based strategy to address x-ray compatibility challenges in advanced packaging metrology. Https://Doi.Org/10.1117/12.3060349, 13426, 865. https://doi.org/10.1117/12.3060349
Zhou, B., Qiao, J., Huang, W., & Li, L. (2025). Defect in-situ monitoring of bump during the bonding process of 3D packaging based on active thermography. Journal of Manufacturing Processes, 154, 1–12. https://doi.org/10.1016/J.JMAPRO.2025.09.065