LOW-COST HYBRID VISION SYSTEM FOR AUTOMATED DIMENSIONAL METROLOGY: MACHINE LEARNING-DRIVEN SMARTPHONE INTEGRATION IN ROBOTIC QUALITY CONTROL APPLICATIONS

Alexandru MOLDOVAN, David CHERECHES, Györk Attila KRASZNAI, Mircea MURAR, Mircea FULEA, Grigore POP, Bogdan MOCAN

Abstract


Traditional dimensional metrology systems require significant capital investment, limiting accessibility for small-scale operations. This study investigates a novel hybrid system combining industrial robotics with consumer-grade imaging technology and pre-trained machine learning models for automated dimensional quality control. A systematic experimental framework utilized a FANUC industrial robot for part positioning relative to a fixed Samsung Galaxy S24 Plus smartphone camera for image acquisition, and YOLOv8 object detection model fine-tuned for dimensional feature extraction. The system's performance was evaluated through repeated measurements on standardized specimens, with statistical analysis of repeatability, reproducibility, and accuracy. Results demonstrated measurement uncertainties within ±0.05 mm for 10÷100 mm features, repeatability coefficients below 2% across 500 cycles, and robot positioning consistency with ±0.02 mm standard deviation. The YOLOv8-based system achieved 98.9% detection accuracy while reducing costs by 78% compared to conventional solutions. The system satisfies ISO 14253-1 requirements for non-critical verification, presenting a viable alternative for resource-constrained environments including educational institutions, prototyping facilities, and SMEs.


Full Text:

PDF

References


Weckenmann, A., et al. (2020). "Coordinate metrology in production engineering." CIRP Annals, 69(2), 681-704.

Hocken, R.J., Pereira, P.H. (2019). Coordinate Measuring Machines and Systems. CRC Press.

Mocan, B., Fulea, M., Olaru, M. and Buchmüller, M., “From Intuitive Programming of Robotic Systems to Business Sustainability of Manufacturing SMEs”. Amfiteatru Economic, 18(41), pp. 215-231, 2016

Samsung Electronics. (2024). "Galaxy S24 Plus Technical Specifications and Imaging Capabilities." Samsung Developer Documentation.

Ultralytics. (2024). "YOLOv8: State-of-the-art object detection and image segmentation." arXiv preprint arXiv:2305.09972.

Chen, L., Rodriguez, M., Kim, S. (2024). "YOLOv8 applications in semiconductor defect detection: Performance analysis and industrial validation." Journal of Manufacturing Systems, 71, 234-247.

Rodriguez, C., Martinez, F. (2024). "Automotive component dimensional measurement using YOLOv8 and smartphone cameras." Measurement Science and Technology, 35(4), 045007.

Jocher, G., et al. (2024). "Transfer learning strategies for YOLOv8 in industrial applications." Computer Vision and Pattern Recognition, 156, 112-125.

Ultralytics. (2024). "YOLOv8 Model Variants: Performance Analysis and Selection Guidelines." Technical Report.

Ashourpour, M., Azizpour, G., & Johansen, K. (2023). “Real-Time Defect and Object Detection in Assembly Line: A Case for In-Line Quality Inspection”. Book Chapter. https://doi.org/10.1007/978-3-031-38241-3_12

Miao, F., Huang, Y., Qian, W., Shi, F., & Lu, Z. (2023). “Intelligent Surface Defect Detection of Hot Rolled Strip Steel Using YOLOv8 Framework”. Proceedings Article. https://doi.org/10.1109/icct59356.2023.10419492

Tao, Q., Chen, Y., & Chen, H. (2023). “A Detection Approach for Wafer Detect in Industrial Manufacturing based on YOLOv8”. Journal Article. https://doi.org/10.1109/safeprocess58597.2023.10295689

Qualcomm Technologies. (2024). "Snapdragon 8 Gen 3: AI and Camera Processing Capabilities." Technical Specification Document.

Samsung Semiconductor. (2024). "NPU Performance Characteristics in Galaxy S24 Series." Technical White Paper.

Kim, J., Park, S., Lee, H. (2024). "Precision measurement capabilities of Samsung Galaxy S24 camera systems in manufacturing applications." Optics and Lasers in Engineering, 189, 107856.

Bințințan, V., Calborean, A., Mocan, M., Macavei, S., Cordoș, A., Ciuce, C., Bințințan, A., Chira, R., Nagy, G., Surlin, V., Timofte, D., Nickel, F., Mueller, B., Dindelegan, G., Ciuce, C., Brad, S., Murar, M., Mocan, B., “New inductive proximity sensor platform for precise localization of small colorectal tumors” , Materials Science and Engineering: C -MATERIALS FOR BIOLOGICAL APPLICATIONS, Volume 106, Article 110146, 2020, ISSN 0928-4931

Zhang, H., Liu, Q., Wang, X. (2024). "Comparative analysis of pre-trained versus custom CNN development for industrial metrology." IEEE Transactions on Industrial Informatics, 20(3), 1847-1856.

Rodriguez, C., Martinez, F., Lopez, J. (2020). "Machine learning-enhanced robotic inspection systems: A comprehensive review." Robotics and Computer-Integrated Manufacturing, 62, 101875.

Schmidt, J., et al. (2021). "Deep learning approaches for automatic hand-eye calibration in robotic systems." IEEE Trans. Automation Science and Engineering, 18(2), 892-903.

Chen, X., et al. (2022). "Reinforcement learning for adaptive robotic metrology systems." IEEE Trans. Industrial Informatics, 18(4), 2847-2856.

Mocan, B., Fulea, M., Brad, S.” Designing a Multimodal Human-Robot Interaction Interface for an Industrial Robot” ADVANCES IN ROBOT DESIGN AND INTELLIGENT CONTROL, V. 371, pp. 255-263, SPRINGER-VERLAG BERLIN

Mocan, B., Fulea, M., Brad, E. and Brad, S., “State-of-the-Art and Proposals on Reducing Energy Consumption in the Case of Industrial Robotic Systems”, Proceedings of the 2014 International Conference on Production Research – Regional Conference Africa, Europe and the Middle East; 3rd International Conference on Quality and Innovation in Engineering and Management, Cluj-Napoca, Romania, 1-5 July, ISBN: 978-973-662-978-5, pp. 328-334, 2014


Refbacks

  • There are currently no refbacks.


JOURNAL INDEXED IN :