Automatic Road Distress Detection and Analysis

Pavement distresses (damages) are important information for evaluating the road condition and conducting the necessary road maintenance activities. Conventional human visual pavement distress detection method is time consuming, very expensive, labour-intensive and dangerous due to exposure to traffic. Numbers of methods have been introduced for road damage detection in the context of fine structure extraction. Due to the unceasing traffic increase, the automation of pavement surface distress monitoring is more and more required. Several techniques are developed for this purpose. But all those approaches access the road condition based on cracks on roads. Due to the various climatic factors, Indian roads suffer from some other types of distresses like potholes also. The goal of this paper is to introduce a novel technique to determine the road condition based on the cracks and potholes on road surfaces. Thus a fully integrated system is proposed for the automatic detection and characterization of distresses in road and flexible pavement surfaces and to detect its severity. The main tasks involved are Collection of images, Distress Detection and Classification using Supervised training approach, Assignment of crack's severity levels to analyze the road performance.