Fault Detection in Steel Belts of Tires Using Magnetic Sensors and Different Deep Learning Models

Tire failures pose significant safety risks, necessitating advanced inspection techniques. This research investigates the application of magnetic sensors and deep learning for detecting defects in steel belts of the tires. It was aim to develop a robust and accurate fault detection system by measuring magnetic field variations caused by defects. In this study, the magnetic image sensor circuit had been designed and then the images obtained from it have been classified as none, crack, and delamination type steel belt errors. Various deep learning models and their hybrid architectures, were explored and compared. Experimental results demonstrate that all models exhibit strong performance, with the Transformer model achieving the highest accuracy of 96.12%. The developed system offers a potential solution for improving tire safety and reducing maintenance costs in industries.