1.18: Vision-Based Detection of Bridge Damage Captured by Unmanned Aerial Vehicles

Project 1.18 Summary:

Bridge inspection is a vital component of any bridge management strategy of a state DOT.
Significant funds are allocated to keeping the over 600,000 bridges in the U.S. safe. A routine
bridge inspection is the most common type of inspection and is often performed from deck,
ground or water levels or from permanent access structures, if available. Visual inspection is the
predominant approach used in a routine inspection. With visual inspection, only basic tools for
cleaning, probing, sounding, measuring, and visual aids are used. However, according to
research, there can be significant variation in the condition ratings assigned to a structure simply
based on visual inspection. The use of unmanned aerial vehicles (UAVs) has recently been
explored for the use of bridge inspections. UAVs equipped with high resolution or infrared
cameras can be used to scan a bridge taking hundreds of images and essentially building a
navigable 3D model of the bridge. Recent advances in machine learning may be employed to
automatically identify different types of bridge damage. This research project will evaluate the
effectiveness of using more autonomous methods for the collection and analysis of bridge deck
images for the purpose of identifying the type and extent of damage in concrete decks.

Sponsors:
Office of the Assistant Secretary for Research and Technology, University Transportation Centers Program, University of Rhode Island, Department of Transportation, RIDOT, Steere Engineering.

Implementation of Research Outcomes:
This project is in its initial research phase. Implementation of Research outcomes will be reported upon completion of initial research.

Impacts and Benefits of Implementation:
This project is in its initial research phase. Impacts and benefits of the research will be reported after the implementation phase.

Related Links:
Coming Soon

Downloadable Documents