Predicting Defective and Good Tyre Quality Status with Pre-Trained CNN Models

Tyres are a critical organization for the safety, performance and driving comfort of vehicles. The qualities of tires are one of the direct characteristics of the travel experience and have a great impact on the safety of passengers and passengers. In recent years, thanks to developments in image processing and artificial intelligence technologies, the use of image-based methods in the color determination process of tires has increased significantly. Considering these conditions, tire errors were detected in two models to detect these extinguisher tire errors. Tire Classification Dataset was used to make these explanations. SqueezeNet and InceptionV3 pre-trained models were used as the classification model. As a result of the work done, a accuracy rate of 94.1% was achieved with the SqueezeNet model and 94.8% with the InceptionV3 model. It is envisaged that tire temperature, which can be integrated into the systems of the proposed models, can be detected quickly and reliably.