Shruti Sabale,
Shraddha Bhawar,
Dinesh Gujar,
Payal Rajput,
S.P. Palekar,
- Student, Department of Mechanical Engineering, Sanjivani College of Engineering, Kopargaon, Ahmednagar, Maharashtra, India
- Student, Department of Mechanical Engineering, Sanjivani College of Engineering, Kopargaon, Ahmednagar, Maharashtra, India
- Student, Department of Mechanical Engineering, Sanjivani College of Engineering, Kopargaon, Ahmednagar, Maharashtra, India
- Student, Department of Mechanical Engineering, Sanjivani College of Engineering, Kopargaon, Ahmednagar, Maharashtra, India
- Student, Department of Mechanical Engineering, Sanjivani College of Engineering, Kopargaon, Ahmednagar, Maharashtra, India
Abstract
This study explores the design of a vision-based screw detection and orientation system for industrial automation, inspection, and robot disassembly. By integrating machine learning algorithms like region-based convolutional neural networks (R-CNN) with traditional image processing and impedance sensing, the system performs real-time screw presence detection, head type identification, and alignment. Three key technologies—deep learning classification, edge-based geometric analysis, and impedance verification—are integrated into a single modular system. The findings indicate that the proposed hybrid system greatly improves detection accuracy, automates disassembly, and complies with Industry 4.0 demands. In addition to detection and orientation, the system architecture is designed with scalability and adaptability in mind, enabling seamless deployment across diverse manufacturing environments. A high-resolution industrial camera captures continuous image streams, which are pre-processed using noise reduction, contrast enhancement, and normalization techniques to ensure consistent performance under varying lighting conditions. The extracted image features are then processed through the R-CNN framework to localize screws within complex assemblies, even in cluttered or partially occluded scenarios. To enhance robustness, the system incorporates edge-based geometric analysis for precise contour extraction and slot orientation measurement. This approach enables accurate differentiation between various screw head types such as Phillips, flathead, Torx, and hexagonal configurations. The integration of impedance sensing further validates screw engagement status during robotic manipulation, ensuring that the detected screw is physically accessible and correctly aligned before initiating disassembly. This multimodal verification significantly reduces false positives and operational errors. The modular design allows independent calibration of each subsystem—vision processing, classification, geometric analysis, and impedance sensing—thereby facilitating easier maintenance and upgrades. Real-time feedback mechanisms are implemented to support adaptive robotic control, where orientation data is transmitted directly to robotic end-effectors for precise torque application and unscrewing operations.
Keywords: Screw detection, R-CNN, robotic vision, automated disassembly, hough transform, impedance sensor, machine learning, smart manufacturing
[This article belongs to International Journal of Advanced Robotics and Automation Technology ]
Shruti Sabale, Shraddha Bhawar, Dinesh Gujar, Payal Rajput, S.P. Palekar. Design and Development of Screw Detection System : A case study. International Journal of Advanced Robotics and Automation Technology. 2026; 04(01):30-36.
Shruti Sabale, Shraddha Bhawar, Dinesh Gujar, Payal Rajput, S.P. Palekar. Design and Development of Screw Detection System : A case study. International Journal of Advanced Robotics and Automation Technology. 2026; 04(01):30-36. Available from: https://journals.stmjournals.com/ijarat/article=2026/view=244006
References
- Simon Mangold, Christian Steiner, Marco Friedmann, Jürgen Fleischer, 2022, Vision-Based Screw Head Detection for Automated Disassembly for Remanufacturing, Procedia CIRP, vol. 105, pp 1–6. DOI NO – 10.1016/j.procir.2022.02.001
- Gomez, E. V., “Development of Screw Detection and Orientation System,” California State University, 2023.
- Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., and Fei-Fei, L., 2010, “ImageNet: A Large-Scale Hierarchical Image Database,” 2009 IEEE Conference on Computer Vision and Pattern Recognition, Miami, FL, June 20–25, pp. 248–255.
- Liu, S., Qi, L., Qin, H., Shi, J., and Jia, J., 2018, “Path Aggregation Network for Instance Segmentation,” Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, June 19–21, pp. 8759–8768.
- Hashemi V., Chen M., Fang L. Modelling and analysis of aerospace remanufacturing systems with scenario analysis. Int. J. Adv. Manuf. Technol. 2016;87:2135–2151.
- Kuo H, Xu J, Yu C, Yan J. Screw defect detection system based on AI image recognition technology. In2020 International Symposium on Computer, Consumer and Control (IS3C) 2020 Nov 13 (pp. 493–496). IEEE.
- Sabri MN, Jaafar Z. DESIGN AND DEVELOPMENT OF SORTING SCREW MACHINE FOR TOOLING WORKSHOP. Jurnal Kejuruteraan, Teknologi and Sains Sosial (JKTSS). 2026 Jan 2;12(1):162–79.
- Hsu WC, Chen HC, Hsu PY, Siao YJ. Research on screw detection using artificial intelligence technology. InInternational Conference on Advanced Information Networking and Applications 2024 Apr 9 (pp. 37–45). Cham: Springer Nature Switzerland.
- Feng GH, Pan YL. Establishing a cost-effective sensing system and signal processing method to diagnose preload levels of ball screws. Mechanical Systems and Signal Processing. 2012 Apr 1;28:78–88.
- Lin WJ, Chen JW, Young HT, Hung CL, Li KM. Developing the smart sorting screw system based on deep learning approaches. Applied Sciences. 2021 Oct 19;11(20):9751.
| Volume | 04 |
| Issue | 01 |
| Received | 13/01/2026 |
| Accepted | 25/02/2026 |
| Published | 11/03/2026 |
| Publication Time | 57 Days |
Login
PlumX Metrics