Automated System for Concrete Damage Classification Identification Using Naïve-Bayesian Classifier

Abstract

The manual inspections of concrete have some limitations. The findings are subjective, manual data processing is expensive and time-consuming, and a range of safety risks are involved with operating at such heights and under heavy traffic. This research aims to study machine learning to detect damage on the concrete surface in MATLAB by using a Naïve-Bayesian classifier. Specifically, the strong point of this automated system is saving cost, time, and high safety. This system can save cost and time because it can detect numerous clear images of the concrete damage type covered in this study at one time. Besides, inspection work can continue even if the expertise is sick, and human emotions do not influence final accuracy. This automated system starts with feature extraction using the Gray Level Co-occurrence Matrix (GLCM) method, followed by the Naïve-Bayes classifier in the classification learned in MATLAB apps. This study aims to develop an automated system to classify the type of concrete damage image classification using Naïve-Bayesian classifier correctly and reduce time during inspection, analysis, and classification. The testing dataset results show that the GLCM method and the Naïve-Bayesian classifier can successfully detect the concrete damage image. Therefore, the created model might be a helpful tool for building management agencies and construction engineers in structure maintenance work.