ABSTRACT
Bridges inspection is the process of implementing a damage detection and characterization strategy for bridges. Few major bottlenecks currently exist that severely limit the effectiveness of existing bridge health management methods, like the requirement of a continuous power supply for each bridge, sensors installation cost …etc. This paper aims to resolve the aforementioned major bottlenecks by combining visual inspection with vibration based methods in a system called “Fly-by Bridges Inspection”. In this system, the inspection truck will carry a swarm of drones, drone charging and communication pads, on-board computers and wireless antenna. The truck will stop beside the target bridge. A swarms of mosquito drones equipped with sticky mechanism and accelerometers will fly to stick with the bridge girder to record the bridge accelerations which needed to assess the bridge condition. Another drone with high resolution camera will fly to detect any existed crack using Faster Convolutional Neural Networks (Faster R-CNN) image processing technique. After finishing the inspection process, drones will back to the inspection truck to download data, charge itself, and be ready for new inspection. This paper will go through all the procedures that make this system success.
