Concrete Bridge Deck Surface Characterization: Resolving Cracks Using Optical Sensors and Image Processing

Concrete bridge decks are a critical structural component in most bridges and the characterization of their deteriorating condition remains as an ongoing challenge for the Department of Transportation. These challenges include lane closures, inspector safety, and subjective results. Visual inspection remains as the primary and most common nondestructive evaluation technique for bridge decks, but recently there has been an increased interest in enhancing or even replacing this method with digital image processing techniques. An automated system would not only increase inspector safety but also produce reliable and repeatable results.

This thesis presents findings pertaining to the evaluation of non-contact measurement techniques in their use in quantifying cracks. In particular, the research compares promising image acquisition systems and analysis techniques. Images from both a controlled laboratory and outdoor environment were used to test the image acquisition systems and image processing methodologies, and establish a measure of both accuracy and efficiency. The outcome of the investigation provides confidence to the feasibility of the automated image processing approach, but also creates a foundation for the application to more complex environments.