Road Defect Classification Using Gray Level Co-Occurrence Matrix (GLCM) and Radial Basis Function (RBF)

Abstract

The road is an important infrastructure, so it is necessary to maintain the road periodically. Currently, the road defect assessment is still manual. Unfortunately, this method takes a long time and can cause ambiguity because of the subjectivity factor. Along with the development of science on image processing technology and machine learning, assessment of road defects can be done automatically by the machine. Road defect classification is the first step in automated road assessment. The image of road defect will be taken from the machine, taking on the features of each defect and classifying the image of its features. GLCM is a feature extract method that has been widely used for image processing. This study classifies some types of road defects, ie potholes, cracks and other defects using the Gray Level Co-occurrence Matrix (GLCM) as a feature extract, while Radial Basis Function (RBF) as an object classification. The proposed method can classify defects with an average of 93% accuracy, 93% precision and 100% recall.