Dataset for Drone-based Inspection of Road Pavement Structures for Cracks

Published: 8 November 2022| Version 1 | DOI: 10.17632/csd32bm8zx.1


The dataset is available online as a benchmarking dataset for drone-based inspection of pavement structures. The data were acquired in an experimental road with a length of 386 meters, belonging to Montmorency Forest laboratory of Université Laval, located in North of Quebec City, on 2021/06/16. The road is mainly used for testing pavement paints, laying techniques, and inspection simulations. A DJI MINI 2 drone was employed to collect images for this dataset. The drone has a 12 megapixels camera with an 83 degrees field of view, capturing 1920 x 1080 images in the continuous recording mode. The drone performs six passes on an experimental road at different altitudes and horizontal speeds. The drone was controlled manually, and the footage was acquired using the embedded camera that stabilized and controlled using the drone's gimbal. Moreover, after data acquisition, the length and width of some cracks and road landmarks were measured for evaluating crack characterization. **** In case of any use, please cite this dataset and our paper ****



Universite Laval


Non-Destructive Testing, Crack, Deep Learning, Drone (Aircraft)