Formulation of Pattern Recognition Framework - Analysis and Detection of Tyre Cracks Utilizing Integrated Texture Features and Ensemble Learning Methods

Abstract

For a safe drive with a vehicle and better tyre life, it is important to regularly monitor the tyre damages to diagnose its condition and chose appropri- ate solution. This paper proposes a framework based on pattern recognition utilizing the strength of texture attributes and ensemble learning to detect the damages on the tyre surfaces. In this paper, a concatenation of the statistical and edge response based texture features derived from Gray Level Co-occurrence Matrix and Local directional pattern are proposed to describe and represent the tyre surface characteristics and their variations due to any damages. The derived fea- tures are provided to train machine learning algorithms using ensemble learning methods for a better under- standing to discriminate the tyre surfaces into normal or damaged. The experiments of tyre surface classifica- tion were conducted on the tyre surface images acquired from Kaggle tyre dataset. The results demonstrated the ability of the combined texture features and ensemble learning methods in effectively analysing the tyre sur- faces and discriminate them with better performance provided by adaboost and histogram gradient boosting methods.