Unmanned Aerial Vehicle Assisted Bridge Crack Severity Inspection Using Edge Detection Methods

Currently, cracks in bridge is a major problem for the existence of human life. Bridge accident increases due to lack of the surveillance of bridges. The cracks are common in bridges due to frequent running of heavy motor vehicles, natural calamities, etc. However, if the crack width increases it may create problem in transportation system. So, it is essential to regularly monitor the critical positions of bridge by capturing their images through unmanned aerial vehicles (UAVs), surveillance camera, vibration sensors, etc. The captured images are processed to monitor the cracks. In this paper, a UAV based bridge crack inspection model focusing on edge detection (ED) mechanisms is proposed to support the Internet of Things (IoTs) applications. ED refers to the process of identification of edge pixels in an image. The images after taken from the UAVs are transmitted to an Information Center (IC) through internet for processing using ED methods to identify the crack. If the crack width is greater than a threshold then actions are taken for bridge crack (BC) repair or bridge reconstruction. In this methodology, the main work is focused on several BCs using different existing standard ED methods such as sobel, prewitt, roberts, Laplacian of Gaussian (LoG), morphological and canny methods. These methods are implemented in MATLAB R2015b and evaluated using Peak Signal to Noise Ratio (PSNR) and Signal to Noise Ratio (SNR).