Speed breaker detection using GLCM features

Abstract

Road accidents are increasing worldwide, that leads to death, injuries and vehicle damages. Most of the accidents happen due to the improper warning sign and unnoticeable speed breakers on the road especially during night. Identification and notification of road signs and speed breakers to the driver at proper time is very important to avoid accidents. In this paper, speed breaker identification using Gray Level Co-occurrence Matrix (GLCM) features is proposed. This method has three stages namely pre-processing, feature extraction and classification. Noise removal, Resizing the image and gray scale conversion has been done as a part of pre-processing. In the feature extraction step, the spatial relationship between the pixels is obtained. GLCM features are the second order statistical features of the image. These features includes correlation, Angular Second Moment, Entropy, Homogeneity and contrast. In this paper, features are consider as the shape, texture and feature statistics. Neural Network based classifier is used in the third stage to identify the presence of speed breaker. The performance of the classifier is evaluated by calculating the confusion matrix.