TireEye: Optical On-board Tire Wear Detection

Abstract

Automotive tire tread depth significantly influences a car’s safety and must therefore be closely monitored. However, there is currently no on-board solution that can measure tire wear with an error of less than 0.6 mm in real-world conditions. This corresponds to 37.5 % of the mandatory minimum tread depth in most countries. In this paper we present the concept of TireEye, which is an optical device mounted inside the wheel well and facing the road. This device records the cross-section of the longitudinal tread groove and extracts its outline by using adaptive canny edge detection. The tread wear indicators serve to calibrate the scale since they provide the smallest allowed tread depth. We validate this technology with several tires, various lighting conditions, and different road surfaces. It provides a mean absolute error of 0.57 mm in real-world conditions, which outperforms all other on-board tire wear detection methods displayed in state-of-the-art. Although the results are very promising, the hardware costs and the susceptibility to dirt might make it difficult for automotive companies to deploy. This can be counteracted with additional use cases like tire pressure estimation, tire damage detection, and road friction coefficient estimation.