Date of Award
Master of Science
Department of Systems Engineering and Management
Vhance V. Valencia, PhD.
Efficient, reliable data is necessary to make informed decisions on how to best manage aging road assets. This research explores a new method to automate the collection, processing, and analysis of transportation networks using Unmanned Aerial Vehicles and Computer Vision technology. While there are current methodologies to accomplish road assessment manually and semi-autonomously, this research is a proof of concept to obtain the road assessment faster and cheaper with a vision for little to no human interaction required. This research evaluates the strengths of applying UAV technology to pavement assessments and identifies where further work is needed. Furthermore, it validates using UAVs as a viable way forward for collecting pavement information to aid asset managers in sustaining aging road assets. The system was able to capture road photos suitable for semi automated Pavement Condition Index (PCI) processing, however the algorithm resulted in a maximum F-Measure of 40%. This result is low and indicates the algorithm is not sufficient for fully automated PCI classification. Accurately detecting road defects using computer vision remains a challenging problem for future research. However, using Autonomous UAVs to collect the data is a viable avenue for data collection, theoretically faster than current methods at freeway speeds.
DTIC Accession Number
Grandsaert, Patrick J., "Integrating Pavement Crack Detection and Analysis Using Autonomous Unmanned Aerial Vehicle Imagery" (2015). Theses and Dissertations. 147.