Setting the Flight Parameters of an Unmanned Aircraft for Distress Detection on the Concrete Runway

Abstract

The airport pavement is annually inspected by a visual survey for the presence of distress to keep a high level of safe air traffic. The basic airport pavement distress is a crack, whose main criterion for evaluation is its width. In relation to air traffic safety, the cracks are divided into small, medium, or large categories according to the severity. A modern way of conduction of airport pavement inspection is the use of unmanned aircraft. This article explores the effect of unmanned aircraft flight parameter settings to recognize the distress on the concrete runway. The image data were obtained from the flights at several altitudes above the runway of a former military airport and processed using commercial multi-view reconstruction software. The output orthomosaic images were evaluated according to pixel resolution, ground sampling distance, and ground resolved distance. The correlation between the variables was statistically analyzed using linear and polynomial regression. The low flight altitudes bring a higher level of captured detail, but only the small area is captured, and more data needs to be processed, whereas higher flight altitudes cover a larger area with less data, but the captured detail is getting low. The findings reveal in order to capture the required detail; it is necessary to divide the length of the runway into individual legs, which allows the flight in lower altitudes above ground level. For the used unmanned aircraft, a balance between the flight altitude and the captured detail corresponding to the flight time per battery has been found. The captured data will be further used to create a database of individual distress for deep learning.