Manoeuvring drone (Tello Talent) using eye gaze and or fingers gestures
The project aims to combine hands and eyes to control a Tello Talent drone based on computer vision, machine learning and an eye tracking device for gaze detection and interaction. The main purpose of this project is gaming, experimental and educational for next coming generation, in addition it is...
| Autor: | |
|---|---|
| Tipo de recurso: | tesis de maestría |
| Fecha de publicación: | 2023 |
| País: | España |
| Institución: | Universitat Politècnica de Catalunya (UPC) |
| Repositorio: | UPCommons. Portal del coneixement obert de la UPC |
| Idioma: | inglés |
| OAI Identifier: | oai:upcommons.upc.edu:2117/395857 |
| Acceso en línea: | https://hdl.handle.net/2117/395857 |
| Access Level: | acceso abierto |
| Palabra clave: | Gesture recognition Drons Eye tracking Avions no tripulats Àrees temàtiques de la UPC::Enginyeria de la telecomunicació::Processament del senyal::Reconeixement de formes |
| Sumario: | The project aims to combine hands and eyes to control a Tello Talent drone based on computer vision, machine learning and an eye tracking device for gaze detection and interaction. The main purpose of this project is gaming, experimental and educational for next coming generation, in addition it is very useful for the peoples who cannot use their hands, they can maneuver the drone by their eyes movement, and hopefully this will bring them some fun. The idea of this project is inspired by the progress and development in the innovative technologies such as machine learning, computer vision and object detection that offer a large field of applications which can be used in diverse domains, there are many researcher are improving, instructing and innovating the new intelligent manner for controlling the drones by combining computer vision, machine learning, artificial intelligent, etc. This project can help anyone even the people who they don¿t have any prior knowledge of programming or Computer Vision or theory of eye tracking system, they learn the basic knowledge of drone concept, object detection, programing, and integrating different hardware and software involved, then playing. As a final objective, they can able to build simple application that can control the drones by using movements of hands, eyes or both, during the practice they should take in consideration the operating condition and safety required by the manufacturers of drones and eye tracking device. The concept of Tello Talent drone is based on a series of features, functions and scripts which are already been developed, embedded in autopilot memories and are accessible by users via an SDK protocol. The SDK is used as an easy guide to developing simple and complex applications; it allows the user to develop several flying mission programs. There are different experiments were studied for checking which scenario is better in detecting the hands movement and exploring the keys points in real-time with low computing power computer. As a result, I find that the Google artificial intelligent research group offers an open source platform dedicated for developing this application; the platform is called MediaPipe based on customizable machine learning solution for live streaming video. In this project the MediaPipe and the eye tracking module are the fundamental tools for developing and realizing the application. |
|---|