Pavement Distress Detection, Classification, and Analysis Using Machine Learning Algorithms: A Survey

Abstract

Distress is any observable deterioration or damage that negatively impacts the road’s performance and safety. Potholes cracks, rutting, and bleeding are a few examples of distress. Maintaining the roads and detecting distress on the surface of the road is critical to avoid impending accidents, consequently saving lives. The article primarily explains the systematic approach of autonomous techniques for detecting distress such as potholes and cracks. Among the array of methods employed for finding distress, the current study reviews the features of three different artificial intelligence (AI) techniques, which include machine and deep learning approaches. Applications of these techniques help in finding pavement distress apart from the vibration, 2D, and 3D methods. This systematic approach explains the autonomous techniques for detecting surface distress, the scope of combining those approaches, and their limitations. Furthermore, the review helps the researchers to widen their knowledge about the various methods in use. It also offers details about the available datasets for experimentation to establish smart cities and transportation.

Keywords

• 2D and 3D methods for road images
• automated road monitoring
• pavement distress
• pothole detection
• road surface cracks and damages
• and vibration methods