Detection of Tyre Expiration using Convolutional Neural Network and Optical Character Recognition
Tyre expiration detection is an important aspect of road safety and vehicle maintenance because aged or expired tyres may experience deterioration that can increase the risk of blowouts, reduced traction, and compromised vehicle handling. Conventional methods of detecting tyre expiration dates largely depend on manual inspection, which can be time-consuming, prone to human error, and affected by variations in human judgment. In addition, expiration dates printed on tyre sidewalls can be difficult to read because of low contrast, text curvature, varying lighting conditions, image distortion, and other markings or debris. Therefore, there is a need for an automated and reliable approach for detecting and recognizing tyre expiration dates. The aim of this study was to develop an automated system for detecting and recognizing tyre expiration dates using Convolutional Neural Network (CNN) and Optical Character Recognition (OCR) techniques. The study specifically sought to improve the readability of expiration-date markings under challenging environmental conditions and evaluate the performance of the developed system against existing work. The methodology involved the collection of tyre sidewall images containing expiration-date markings from different environments, including gas stations and parking lots. Image preprocessing techniques, including contrast enhancement, noise reduction, grayscale conversion, thresholding, blurring, text segmentation, and text-region extraction, were applied to improve image quality and readability. A CNN-based approach was employed for text detection and localization, while Tesseract OCR was integrated to recognize and extract the expiration-date characters from the detected regions. The developed system was subsequently tested under varying lighting and image conditions, and its performance was evaluated using recognition accuracy and other performance measures. The results showed that the developed CNN-OCR-based system achieved an average accuracy of 92.0% in detecting and recognizing tyre expiration dates. The findings demonstrate that the integration of image preprocessing, CNN-based detection, and OCR provides a feasible approach for automating tyre expiration detection in real-world environments. The system can improve the efficiency and reliability of tyre inspection, thereby supporting proactive vehicle maintenance and contributing to enhanced road safety. Further improvement is recommended to increase recognition performance under more challenging real-world conditions.
Download Paper