Tire Quality Inspection System Based on Deep Learning

Abstract

There have been many reports of accidents caused by the use of damaged and worn tires, and these accidents are more common on highways and during the rainy season. Although this is a common problem, many people cannot distinguish good tires from worn ones, increasing the risk of ending up with good tires on the road. A few years ago, the main technology for checking tire size was manual inspection. An important method is to determine the grade of the tread pattern bed by checking the depth and the shoulder pattern bed. Tires are one of the most important parts of a vehicle as they actively support driving. However, they often disagree when it comes to proper inspection and maintenance. Most of the time, the general public apparently does not care about their tires. Many will experience tooth wear and flank damage, and failure to follow up on these problems will cause long-term damage. However, this method is too expensive to use in a family car. This article presents a model used as a running image that can distinguish broken tires from rubbing tires. The model is based on the image displayed outside the user-supplied tire and determines its status after comparing it with the model data using the deep learning algorithm ResNet50. This model is made to remind you that it can be used in addition to equipment suitable for use in real life applications. With regulation by regulatory agencies, tire accidents can be reduced and damage to people and property on public roads can be prevented.