Road Surface Types Classification Using Combination of K-Nearest Neighbor and Na?ve Bayes Based on GLCM

摘要

The automatic capability of determining the road surface type is essentialinformation for autonomous vehicle navigation such as wheelchair and smartcar. This factor is crucial because determining the type of road surface canincrease security for auto vehicle users. This study used texture information toextract features from pictures using Gray Level Co-occurrence Matrix (GLCM),and combine K-Nearest Neighbor classifier (KNN) and Na?ve Bayes classifier(NB) to characterize surface objects into three road classes, i.e., asphalt, gravel,and pavement. The combination of 2 classification methods is then written asKNB. The classification performance of KNB will compare with anotherclassifier. In this study, there were 750 images of original roads (asphalt, gravel,and Pavement) that were arranged into a dataset. The results show that theclassification accuracy using KNB is higher than the comparison classificationmethods.