Geolocation and counting of people with aerial thermal imaging for rescue purposes

Abstract

Thermography has become more frequently used in rescue operations when used together with flight technologies such as unmanned aerial vehicles (UAVs). This is due to its non-invasive and powerful supervision characteristics in the spectrum range not perceivable for the human eye or for a standard camera. This paper presents a developed system based on the synergy of a UAV and a counting and geolocation algorithm that detects people with aerial shots in areas of difficult access. The system integrates a thermal camera to a UAV, thus being useful in different scenarios such as floods, fires or wooded areas. For this purpose, the UAV navigation paths are configured from an earth station using a telemetry-specific software. Thermal images will be recorded during the mission at a height determined by the operator, which will later be processed to filter, discriminate, count and geolocalize people at risk. The processing of the images is done by means of artificial vision tools combined with Artificial Neural Networks (ANN).