Road surface classification based on LBP and GLCM features using kNN classifier

Autonomous ground vehicle (UGV) technology has shown a fast development this past year and proven to be useful. The use of UGV technology is restricted on a particular road condition. Classification of the road is an essential process in UGV, especially to control the autonomous vehicle. For example, the speed could be adjusted by referring to the road type, these processes require a fast-computational time. This research focuses on finding the most discriminant feature while keeping the number of features into a minimum to obtain fast computational time and accurate classification result. One can experiences difficulties because the condition of the road varies, this research proposes a combination of gray level co-occurrence matrix (GLCM) a statistical method to extract feature and local binary pattern (LBP) feature to improve the robustness of the features. The kNN classifier is used to do the classification with the accuracy of 98% and 12 picture processed per second