Vision-based identification of tire inflation pressure using Tire-YOLO and deflection

Abstract

Tire inflation pressure has a significant impact on vehicle performance and safety. Existing tire pressure identification methods rely on sensors, which have limitations such as high maintenance costs and poor durability. This study proposes a non-contact method for tire pressure identification using Tire-YOLO and deflection. Firstly, a Tire-YOLO model is developed to identify tire specifications in images, which are used to obtain recommended pressure referring to the standards. Then, the statistical relationship between actual and recommended pressures is established to obtain statistical pressure. Next, the statistical pressure is corrected using the visually identified tire deflection, serving as an approximation of actual pressure. Finally, field tests are conducted on different tire types to verify the proposed method. The results indicate that the inflation pressures determined using this method are within 10% of actual pressures for all the cases. The proposed method provides a potential way for rapid identification of inflation pressure.