Application of drones and image processing for bridge inspections in south africa

Abstract

In South Africa, visual inspections for all bridge structures are required every five years. A need was identified to establish if it would be practical to utilise technologies, such as drones and photogrammetry, to improve inspections and to reduce the cost of inspections. Drone and photogrammetry technologies are large components in the Fourth Industrial Revolution (4IR). For proof of concept, two bridges were inspected using a drone. The drone inspection data captured for the two bridges was processed and point cloud models were created. These point cloud models were then used to establish whether defects, identified on these bridges previously by way of regular TMH 19 visual assessments, could be identified from the point cloud models. The outcome of the study showed that it would be possible to use processed images from drone inspections or other image capturing methods to inspect bridges and for monitoring purposes. The use of these technologies could improve the inspection methodology, not only of individual structures, but also of a network of structures as required by road authorities. Recommendations are made for future work.