Automated Tyre Health Assessment via EfficientNetB0 for Safer Transportation Networks

The nature of tires is of great importance to the safety of the vehicle, its performance and efficiency because the defect of tires contributes significantly to the safety of roads and the performance of the vehicle. Early identification of such kind of defects may help to avoid disasters, reduce downtime and save on maintenance costs. According to the EfficientNetB0 model, a modern convolutional neural network (CNN) that is characterised by its cost-effective nature and excellent performance, this paper provides an automated method of classifying tyres in terms of quality. According to the visual analysis of the digital tire image, the process will classify tires as defective or in good condition. The training, validation and model testing are based on picture collection that was taken in 1854 of the open-source Kaggle site. The EfficientNetB0 model is specifically custom-built to work with tire quality classification. The findings indicate that the accuracy in differentiating between the good tires and bad tires is quite high as the 96.72% classification accuracy of the model indicates. The paper focuses on the opportunities of deep learning technique, specifically transfer learning and fine-tuning, to automate tyre inspection processes and, in connection with this, minimise human labour requirements, thereby making vehicle inspection more efficient and roads safer to use. The paper presents how practical deep learning is in practice and how a scalable and efficient approach may be deployed to the automotive industry in the context of real-time tire quality inspection. The effort will help to achieve SDG objectives in Decent effort and Economic Growth, Industry Innovation and Infrastructure, Sustainable Cities and Communities, Responsible Consumption and Production, Climate Action and Life on Land.