This thesis focuses on developing computer vision algorithms capable of extracting tyre-related information from lateral onboard vehicle cameras. By analysing visible tread patterns, the system aims to classify tyre type (e.g., summer, winter, performance) and estimate wear level under real driving conditions. The research will involve image preprocessing, deep learning classification models, robustness evaluation under varying lighting and motion conditions, and real-time feasibility analysis.
The work will be conducted in cooperation with a car manufacturer, providing real vehicle camera datasets.
