01768067

Component

https://doi.org/10.1177/0361198121999605

This paper presents a framework to better identify and measure defects in a bridge using drone-based inspection images integrated with grayscale image enhancement techniques. For this study, a DJI Matrice 210 drone was used for the inspection of a three-span timber bridge with concrete decking located in Keystone, South Dakota. During the inspection, the drone recorded a series of videos of the bridge using the MOVie (MOV, video file extension) video format. MOV-based image analysis was conducted to identify a variety of defect types (i.e., efflorescence, water leakage, spalling, and discoloration) on the bridge. For improvement of defect visibility, the grayscale image enhancement technique was applied to determine visually enhanced images for the individual defect. The technique used grayscale image histogram processing that can adjust images using realignment of contrast histograms, in which contrasts of each pixel of the grayscale images have their own number from 0 for black to 255 for white in the image. With the enhanced images, pixel-based measurement was conducted to quantify the defects, including efflorescence (3.75?m2), water leakage (4.21?m2), spalling (0.74?m2), and discoloration (2.12?m2). Based on these findings, the grayscale drone inspection image enhancement technique enabled the demonstration of defect visibility adjustment and improvement for more reliable identification and measurement of the defects in the bridge.

The data collected from the bridge inspections that support the findings of this study are available from the corresponding author on reasonable request. © National Academy of Sciences: Transportation Research Board 2021.

English

pp 603-612

2021-8

Transportation Research Record: Journal of the Transportation Research Board

Web

References (28)

Bridges; Defects; Drones; Histograms; Image analysis; Inspection; Measurement

Aviation; Bridges and other structures; Vehicles and Equipment

TRIS, TRB, ATRI

Mar 21 2021 3:10PM