Detection of Corrosion Defects in Steel Bridges by Machine Vision

Existing bridges are critical components of transportation infrastructure manly due to a huge volume of different corrosion. Corrosion reduced the performances of bridges and decrease their life services. Towards automatic detection of corrosion defects during inspections, a novel methodology is here proposed making use of machine vision concepts. Indeed, different types of corrosion can be detected by image processing techniques that can be an appropriate tool also for the prediction of the damage evolution in bridges. Clustering K-means algorithms on image segmentation have been used to classify corrosion defect levels.