Nur Najiha Kamarulzaman, Nor Salwa Damanhuri, Nuraina Husna Mohamad Asri, Nor Azlan Othman, Noor Azlina Mohd Salleh, Belinda Chong Chiew Meng, Anik Nur Handayani, Tomonori Kato
Manual inspection of additive manufacturing is time-consuming and error-prone, making it unsuitable for high-speed production. Real-time automated systems should ensure precision and consistency of defect detection. Hence, this paper presents the development and evaluation of an automated defect classification system for additive manufacturing (AM) products using You Only Look Once version 8 (YOLOv8) and LabVIEW. This study utilized a dataset of 1200 images of AM products from the Malaysia Automotive Robotics & IoT Institute (MARii). YOLOv8, a state-of-the-art object detection technique, was used to develop a defect classification model. Following the development of the classification model, it was implemented on the Karakuri machine by interfacing the hardware with National Instruments' myRIO and LabVIEW. An infra-red sensor triggers image capturing via a USB camera system, while the real-time classification system activates the servo-based sorting mechanisms. Data augmentation techniques were deliberately applied to improve robustness during model training. The system achieved an average accuracy of 93.39 and demonstrated satisfactory performance across all evaluation criteria: precision, recall, and F1-score; thereby confirming its effectiveness in classifying defect and non-defect products. In conclusion, the results validate the designed system as a practically feasible and efficient approach toward automating quality control in additive manufacturing, reducing dependence on manual inspection while improving consistency and operational efficiency. © 2026 The Authors. Published by Universiti Malaysia Pahang Al-Sultan Abdullah Press. This is an open access article under the CC BY-NC 4.0 license
Faculty of Electrical Engineering, Universiti Teknologi MARA, Cawangan Pulau Pinang, Pulau Pinang, Malaysia; Faculty of Mechanical Engineering, Universiti Teknologi MARA, Selangor, Shah Alam, Malaysia; Department of Electrical Engineering and Informatic, Universitas Negeri Malang, Malang, Indonesia; Faculty of Science and Engineering, Department of Mechanical Engineering, Hosei University, Tokyo, Japan
