Object detection and tracking using an UAV

Abstract

This project is enclosed in the field of computer artificial vision and developed by taking profit of the obtained knowledge during the Computer Engineering degree and the computation specialization. Throughout the document we will scrutinize the state of art of the ongoing computational systems, analyzing the history and evolution of them. Owing to develop the final application designed to run in a reduced size and performance system, we will focus on several optimized computational methods, including ad-hoc developed Neural Networks models to serve for this purpose. As a consequence of this analysis, we will develop an application capable of recognizing objects in real time, by using the video streaming from an Unmanned Aerial Vehicle (UAV) which will be running entirely on a medium budget smartphone. As a result of this study, the developed application will be able to automatically command the aircraft to follow a determined target by using the recognitions in the video thus adapting to the position changes of the desired objective.