Automated Vin Identification Approach With Computer Vision For Vehicle Recognition

• Lucas Buligon Antunes UTFPR
• Guilherme Loan Schneider UTFPR
• Igor Wiese UTFPR
• Thiago França Naves UTFPR
• Anderson Soares UFG

DOI: https://doi.org/10.5753/eniac.2025.13666

Abstract

Driven by the need to handle vehicle fraud, which exceeded R$504 million in the first half of 2024 in Brazil, this work presents an automated computer vision system for chassis number detection and recognition. The proposed solution is structured as a five-step pipeline, encompassing the localization of the region of interest using Oriented Bounding Boxes (OBB) and its orientation correction, Optical Character Recognition (OCR), and a final refinement stage through business rule-based post-processing. The main contribution lies in demonstrating that training enrichment with synthetic data is crucial for ensuring system robustness in real-world scenarios, allowing the pipeline to achieve 96% accuracy in full chassis number reading. Future improvements will focus on optimizing the accuracy for perfect reading and exploring end-to-end architectures to simplify the solution and assess its real-time computational performance.

References

American Journal of Transportation (2024). Compliance and regulatory benefits of using vin decoders in international vehicle shipping. [link]. Acessado em 2025-06-19.

Associação Brasileira de Normas Técnicas (2001). NBR 6066:2001 - road vehicles - vehicle identification number (vin).

Confederação Nacional das Seguradoras (2024). Quantificação da fraude no mercado de seguros brasileiro – relatório parcial 2024. [link].

Lubiato, K. (2015). Como as seguradoras combatem as fraudes em automóveis. [link]. Acessado em 2025-08-03.

Patil, A. V. and Dhanvijay, M. M. (2015). Engraved character recognition using computer vision to recognize engine and chassis numbers: Computer vision technique to identify engraved numbers. In 2015 International Conference on Information Processing (ICIP), pages 151–154. IEEE.

Prakash, G., K, A., Maroof, M. M., Maria Fernandes, S., Roopashree, Bekal, A., and Shetty, R. G. (2023). Engraved character recognition using computer vision to recognize engine number with environmental condition setup. In 2023 IEEE International Conference on Distributed Computing, VLSI, Electrical Circuits and Robotics (DISCOVER), pages 1–7.

Ramshankar, Y. and Deivanathan, R. (2018). Development of machine vision system for automatic inspection of vehicle identification number. International Journal of Engineering and Manufacturing (IJEM), 8(2):21–32.

Shah, P., Karamchandani, S., Nadkar, T., Gulechha, N., Koli, K., and Lad, K. (2009). Ocr-based chassis-number recognition using artificial neural networks. In 2009 IEEE International Conference on Vehicular Electronics and Safety (ICVES), pages 31–34.

Souza, L. R. d. S., Oliveira, R. M. M., Stoppa, M. H., and Caldas, J. O. (2014). Desenvolvimento de processo e dispositivo para inspeção da gravação de chassi utilizando visão de máquina. In ABCM Symposium Series in Mechatronics - Vol. 6, pages 1326–1333. ABCM.

Terven, J., Cordova-Esparza, D.-M., Romero-González, J.-A., Ramírez-Pedraza, A., and Chávez-Urbiola, E. A. (2025). A comprehensive survey of loss functions and metrics in deep learning. Artificial Intelligence Review, 58(7).

Yang, C.-H. and Feng, H.-S. (2020). One-stage vehicle engine number recognition system. In 2020 International Conference on Advanced Robotics and Intelligent Systems (ARIS), pages 1–4. IEEE.