A METHOD FOR CRACK EXTRACTION ON CONCRETE SURFACE IMAGES BASED ON NAIVE BAYES FILTER

Abstract

This study proposes a method for extracting cracks from concrete surface images using a naive Bayes filter. The naive Bayes filter, a simple machine learning technique, is used to probabilistically classify crack and non-crack regions by evaluating the size, color, and shape characteristics of regions appearing in binary images. The analysis results demonstrate that the proposed method is capable of extracting crack regions by visualizing the calculated crack probabilities from the naive Bayes filter. Additionally, the proposed method has the potential to improve classification accuracy by estimating the probability density function based on a small amount of training data without the data collection costs.