Laboratory trial of a system for vision based road profile analysis detection using stereo vision

Inspection of bridge structures is an important consideration for governmental bodies as early detection of structural damage can facilitate early repairs, theoretically reducing maintenance costs for underfunded infrastructure management departments. This paper investigates the use of a computer vision-based system for road profile analysis, which can then be used as part of a larger damage detection solution. A brief introduction to traditional methods for this task is presented, followed by a state-of-the-art review of computer vision-based approaches. A stereo vision-based method developed in Queen's University Belfast is then detailed in combination with a laboratory trial to validate the developed concept in addition to verification of validation methods. A discussion of the obtained results followed by recommendations for future work concludes the paper.