Engineering Commons^™

Analysis Of Segmentation Algorithms For Pavement Distress Images, Allen Downey, Haris N. Koutsopoulos, Ibrahim El Sanhouri

Allen B. Downey

Collection and analysis of pavement distress data is an important component of any pavement‐management system. Various systems are currently under development that automate this process. They consist of appropriate hardware for the acquisition of pavement distress images and, in some cases, software for the analysis of the collected data. An important step in the automatic interpretation of images is segmentation, the process of extracting the objects of interest (distresses) from the background. We examine algorithms for segmenting pavement images and evaluate their effectiveness in separating the distresses from the background. The methods examined include the Otsu method, Kittler's method, a …