Adaptive Image Quality Assessment and Enhancement for Robust Automatic Number Plate Recognition in Vehicle Theft Detection

Kamlesh Kumawat; Anubha Jain

doi:10.57159/jcmm.5.3.26825

Authors

Kamlesh Kumawat Department of Computer Science & IT, IIS (Deemed to be University), Jaipur, Rajasthan, India https://orcid.org/0009-0002-1115-776X
Anubha Jain Department of Computer Science & IT, IIS (Deemed to be University), Jaipur, Rajasthan, India https://orcid.org/0000-0002-5785-1266

DOI:

https://doi.org/10.57159/jcmm.5.3.26825

Keywords:

ANPR, EDSR, Optical Character Recognition, Vehicle Theft Detection

Abstract

Vehicle theft remains a worldwide problem, and stolen-vehicle recovery rates stay low despite widespread surveillance, GPS, and smart transport systems. Automatic Number Plate Recognition (ANPR) offers scalable real-time identification but degrades under motion blur, low light, occlusion, and environmental distortion, since most systems presuppose high-quality input. This work proposes a three-stage adaptive pipeline whose contribution lies in the systematic integration of established image quality assessment, enhancement, and recognition components with a calibrated quality-gating mechanism for degraded plate imagery, rather than in new algorithms. Each image is classified by blur severity using the Variance of Laplacian, with empirical thresholds (Not Blurry > 100; Mildly Blurry 20 to 100; Heavily Blurry < 20). Targeted enhancement then follows: bilateral filtering (d = 9, σC = σS = 75) for mild blur and Enhanced Deep Super-Resolution (EDSR ×4, 32 residual blocks, 256 channels, trained on DIV2K) for heavy blur. Detection uses a Haar Cascade, and recognition uses EasyOCR and Tesseract OCR 4.1.1, with recognized plates cross-referenced against a theft database. On the primary dataset, the system reaches 96% (HD), 94% (mild blur), and 91% (heavy blur), attaining 84% on the Kaggle benchmark and 92% on the low-resolution UFPR-ALPR subset. These are gains of 16 and 73 percentage points on the primary dataset, 20 on Kaggle, and 52 on the UFPR-ALPR subset over a conventional baseline. An ablation study confirmed each component's contribution, and the pipeline reaches state-of-the-art accuracy on degraded and low-resolution plate subsets, with applicability to law enforcement and intelligent transportation systems.

References

[1] N. H. F. Syahril and N. N. R. A. Mokobombang, "Enhanced visibility of vehicle license plate recognition systems using adaptive Gabor filters," in Proceedings of the 6th International Conference on Informatics, Multimedia, Cyber and Information System (ICIMCIS), pp. 577–582, 2024.

[2] R. Panahi and I. Gholampour, "Accurate detection and recognition of dirty vehicle plate numbers for high-speed applications," IEEE Transactions on Intelligent Transportation Systems, vol. 18, no. 4, pp. 767–779, 2017.

[3] R. Sridhar, R. A. Suresh, S. Babu, and P. Ravisankar, "Improving low-light image quality for object detection and license plate recognition," Revista Romana de Informatica si Automatica, vol. 35, no. 2, pp. 87–98, 2025.

[4] X. Luo, Y. Huang, and W. Miao, "Real-world license plate image super-resolution via domain-specific degradation modeling," in Proceedings of the 2024 IEEE Conference on Artificial Intelligence (CAI), pp. 1175–1180, 2024.

[5] D. Shah, V. Parwanda, A. Raj, and H. K. Thakkar, "Towards automatic number plate recognition from static images using image enhancement and deep learning," in Proceedings of the 2025 International Conference on Emerging Techniques in Computational Intelligence (ICETCI), 2025.

[6] S. Du, M. Ibrahim, M. Shehata, and W. Badawy, "Automatic license plate recognition (ALPR): A state-of-the-art review," IEEE Transactions on Circuits and Systems for Video Technology, vol. 23, pp. 311–325, Feb. 2013.

[7] A. Krizhevsky, I. Sutskever, and G. E. Hinton, "Imagenet classification with deep convolutional neural networks," in Advances in Neural Information Processing Systems (NeurIPS), pp. 1097–1105, 2012.

[8] J. Redmon and A. Farhadi, "Yolov3: An incremental improvement," arXiv preprint arXiv:1804.02767, 2018.

[9] R. Laroca, E. Severo, L. A. Zanlorensi, L. S. Oliveira, G. R. Gonçalves, W. R. Schwartz, and D. Menotti, "A robust real-time automatic license plate recognition based on the yolo detector," in Proceedings of the International Joint Conference on Neural Networks (IJCNN), pp. 1–10, 2018.

[10] R. Laroca, L. A. Zanlorensi, G. R. Gonçalves, E. Todt, W. R. Schwartz, and D. Menotti, "An efficient and layout-independent automatic license plate recognition system based on the yolo detector," IET Intelligent Transport Systems, vol. 15, no. 4, pp. 483–503, 2021.

[11] S. Zherzdev and A. Gruzdev, "Lprnet: License plate recognition via deep neural networks," arXiv preprint arXiv:1806.10447, 2018.

[12] S. M. Silva and C. R. Jung, "License plate detection and recognition in unconstrained scenarios," in Proceedings of the European Conference on Computer Vision (ECCV), pp. 3–19, 2018.

[13] J. L. Pech-Pacheco, G. Cristóbal, J. Chamorro-Martinez, and J. Fernández-Valdivia, "Diatom autofocusing in brightfield microscopy: A comparative study," in Proceedings of the International Conference on Pattern Recognition (ICPR), vol. 3, pp. 314–317, 2000.

[14] S. Pertuz, D. Puig, and M. A. Garcia, "Analysis of focus measure operators for shape-from-focus," Pattern Recognition, vol. 46, no. 5, pp. 1415–1432, 2013.

[15] M. Bansal, M. Kumar, M. Sachdeva, and A. Mittal, "Transfer learning for image classification using VGG19: Caltech-101 image data set," Journal of Ambient Intelligence and Humanized Computing, vol. 14, pp. 3609–3620, 2023.

[16] C. Tomasi and R. Manduchi, "Bilateral filtering for gray and color images," in Proceedings of the IEEE International Conference on Computer Vision (ICCV), pp. 839–846, 1998.

[17] C. Dong, C. C. Loy, K. He, and X. Tang, "Accelerating the super-resolution convolutional neural network," in Proceedings of the European Conference on Computer Vision (ECCV), pp. 391–407, 2016.

[18] K. He, X. Zhang, S. Ren, and J. Sun, "Deep residual learning for image recognition," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 770–778, 2016.

[19] W. Shi, J. Caballero, F. Huszár, J. Totz, A. P. Aitken, R. Bishop, D. Rueckert, and Z. Wang, "Real-time single image and video super-resolution using an efficient sub-pixel convolutional neural network," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 1874–1883, 2016.

[20] B. Lim, S. Son, H. Kim, S. Nah, and K. M. Lee, "Enhanced deep residual networks for single image super-resolution," in Proceedings of the IEEE CVPR Workshop on New Trends in Image Restoration and Enhancement (NTIRE), pp. 136–144, 2017.

[21] X. Wang, K. Yu, S. Wu, J. Gu, Y. Liu, C. Dong, Y. Qiao, and C. C. Loy, "Esrgan: Enhanced super-resolution generative adversarial networks," in Proceedings of the European Conference on Computer Vision (ECCV) Workshops, pp. 63–79, 2018.

[22] R. M. Nascimento, R. Laroca, J. A. Lambert, W. R. Schwartz, and D. Menotti, "Super-resolution of license plate images using attention modules and a generative adversarial network," Computers & Graphics, vol. 113, pp. 85–96, 2023.

[23] J. Singh and B. Bhushan, "Real time indian license plate detection using deep neural networks and optical character recognition using LSTM Tesseract," in Proceedings of the 2019 International Conference on Computing, Communication, and Intelligent Systems (ICCCIS), pp. 347–352, 2019.

[24] H. Li, P. Wang, and C. Shen, "Toward end-to-end car license plate detection and recognition with deep neural networks," IEEE Transactions on Intelligent Transportation Systems, vol. 20, no. 3, pp. 1126–1136, 2019.