A low-cost video-based pavement distress screening system for low-volume roads

ABSTRACT

This study investigates the use of consumer-grade video camera to develop a low-cost pavement distress screening system, with a particular focus on low-volume roads. Due to the oblique view of video data captured in the car front, existing crack detection algorithms that are tailor-made for nadir view cannot be used directly. Instead, we propose and develop a Video-based PAvement Distress Screening (VPADS) system, which includes an automatic data processing workflow by first defining a Region of Interest (RoI) through robust line fitting of the two side lanes. The pavement condition can be assessed within the RoI by implementing a multi-scale ridge detection filter followed by a boundary contour analysis. Since the proposed VPADS system is designed for preliminary screening purpose, each video image scene is examined if there exists any potential crack or distress feature(s). We tested the proposed system on video data collected for two low-volume roads (Highway 624 and 668) in Northern Ontario region, Canada. The overall accuracy of detecting crack and distress features yielded 80%. Instead of replacing traditional field inspection or high-end multi-sensor pavement evaluation system, the proposed VPADS system aims to provide a computer-aided screening solution for transportation authorities of a vast rural road/highway network and with limited inspection budget. The scope of the work can be further expanded by developing a crowdsensing inspection network built by any regular road user, who can also act as a data provider for the regional/municipal pavement distress screening system.

KEYWORDS:

• crack detection
• canny edge detector
• DBSCAN
• lane detection
• low-volume roads
• pavement distress
• RANSAC
• ridge detection
• surface treated pavement
• video processing

Acknowledgments

The authors would like to thank Mr. Zheng Niu, the former regional GIS coordinator from the MTO – Northeastern Regional Office, who provided the video dataset for experimental testing, Dr. Ningyuan Li, Senior Pavement Management Engineer of MTO, who provided valuable comments on the research, and Dr Qin Zou, Assistant Professor from Wuhan University, who provided the sample pavement images as shown in Figure 4. The research is financially supported by MTO through its Highway Infrastructure Innovation Funding Program (HIIFP) and the NSERC Discovery Grant.