This research develops a platform for investigating computer vision-based bridge inspection algorithms, termed Random Bridge Generator (RBG). The RBG produces synthetic environments with six different types of bridges (slab bridges, beam bridges, girder bridges, arch bridge, cable-stayed bridges, and suspension bridges) and their structural components that are of utmost interest according to the field inspection standards. The synthetic bridges are generated randomly, following relevant design codes and standards. This research demonstrates the effectiveness of the RBG by producing large amount of training data for semantic segmentation of structural components. The produced dataset consists of…. Using the dataset, an algorithm is trained, and promising results are obtained. The authors expect that the developed RBG is a viable tool for enabling the investigation of complex autonomous systems for bridge visual inspections. The Python scripts for RBG is made public at XXXX upon publication.
Updates:
- Oct 2022: ABC
- June 2022: Our paper ABC.
- July 2021: ABC