Corona Discharge Classification Based on UAV Data Acquisition

The Corona discharge represents one of the main concerns of the century in the field of electricity and powerline design, as it is both harmful for the environment as well as very costly to cover for the voltage losses on the transport. In this paper, we propose the design for a measurement and analysis platform for this phenomenon that can estimate the Corona discharge losses based on a video stream coming from an electro-optical sensor placed on board of an Unmanned Aerial Vehicle (UAV). After experimenting on a dataset generated from Corona discharge videos, our results suggest that of the supervised learning algorithms taken into consideration those based on Radial-Basis approaches are the most effective in classifying the information being fed to it. Before the classification task, we propose a starter set of filtering operation, composed of grayscale reduction, mean blurring and binary transformation.