Publication: Reconocimiento de una plataforma para aterrizaje de UAVs mediante procesamiento de imágenes
Loading...
Files
Official URL
Full text at PDC
Publication Date
2017
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Citations
Exportar
Abstract
A lo largo del tiempo y gracias a los grandes avances tecnológicos, el ser humano ha intentado automatizar cada uno de los procesos que intervienen en su día a día con la intención de facilitar la vida de las personas. Uno de estos grandes avances es la posibilidad de enviar vehículos aéreos no tripulados (Unmanned Aerial Vehicle (UAV)) a zonas de difícil acceso para el ser humano, ya sea por razones ambientales, de seguridad o porque suponga un coste económico insostenible. Sin embargo, existen varios problemas que acotan las situaciones en las que este tipo de avances pueden ser aplicados, entre ellos la poca precisión transmitida por la geolocalización actual.
Es por esto que en el presente proyecto se propone una solución al problema de ubicar un UAV en un punto exacto, de manera que realmente se convierta en un vehículo no tripulado y no dependa de forma directa del ser humano. El proyecto donde se circunscribe el desarrollo que se presenta tiene como título \Sistema Autónomo para la Localización y Actuación ante Contaminantes en el Mar (SALACOM, 2013) (DPI2013- 46665-C1)"subvencionado por el Ministerio de Economía, Industria y Competitividad donde se plantea la colaboración entre barcos autónomos y drones. Estos últimos teniendo la necesidad de aterrizar sobre una plataforma situada a bordo de un barco. Por tanto, en este proyecto se plantea el reconocimiento de una plataforma de aterrizaje mediante un sistema de visión por computador basado en el análisis de dos tipos de imágenes, térmicas y de color, captadas por sendos dispositivos, en este caso instalados en un teléfono móvil (smartphone).
La plataforma está específicamente diseñada con tal propósito. En efecto, la misma consta de una base de madera sobre la que se han situado cinco figuras geométricas metálicas de distintos colores. Las figuras geométricas, estratégicamente posicionadas entre sí, configuran una disposición geométrica con nula posibilidad de aparición en los entornos de operación del UAV. Los diferentes colores permiten discernir elementos geométricos con distinta forma espectral. Finalmente, su naturaleza metálica permite distinguir distintas reflectancias térmicas en el espectro visible. Para el reconocimiento se dispone siempre de una imagen en color y otra infrarroja. De esta forma se combinan dos procedimientos que aumentan el grado de precisión en la identificación.
Esto es lo ideal y este es el marco del proyecto, al margen de que estas mismas técnicas se utilicen en otros contextos, no necesariamente en colaboración.
Over time and thanks to the great technological advances, mankind has tried to automate each of the processes that intervene in daily life with the intention of making life easier for people. One of these is the possibility of sending Unmanned Aerial Vehicles (UAV) to areas that are difficult for human beings to access, either for environmental reasons, safety reasons or because they entail an unsustainable economic cost. However, there are several problems that limit the situations in which this type of advances can be applied, such as the little precision transmitted by the current geolocation. For this reason, the present project proposes a solution to the problem of locating a UAV at an exact point, so that it really becomes an unmanned vehicle and doesn't depend directly on the human. The project where the development is presented is entitled \Autonomous System for Localization and Action against Pollutants in the Sea (SALACOM, 2013) (DPI2013-46665-C1)" funded by the Ministry of Economy, Industry and Competitiveness where it raises the collaboration between autonomous ships and drones.The latter needing to land on a platform aboard a ship. Therefore, this project involves the recognition of a landing platform through a computer vision system based on the analysis of two types of images, thermal and color, captured by two devices, in this case installed in a Smartphone. The platform is specifically designed for this purpose. In fact, it consists of a wooden base on which five metallic figures of different colors have been placed. The geometric figures, strategically positioned between them, configure a geometric arrangement with no possibility of appearance in the operating environments of the UAV. The different colors allow geometrical elements with different spectral shapes to be discerned. Finally, its metallic nature allows different thermal reflectances in the visible spectrum to be distinguished. A color and an infrared image are always available for recognition. In this way two procedures are combined that increase the degree of accuracy in identification. This is the ideal scenario and the framework of the project, regardless of whether these same techniques are used in other contexts, not necessarily in collaboration.
Over time and thanks to the great technological advances, mankind has tried to automate each of the processes that intervene in daily life with the intention of making life easier for people. One of these is the possibility of sending Unmanned Aerial Vehicles (UAV) to areas that are difficult for human beings to access, either for environmental reasons, safety reasons or because they entail an unsustainable economic cost. However, there are several problems that limit the situations in which this type of advances can be applied, such as the little precision transmitted by the current geolocation. For this reason, the present project proposes a solution to the problem of locating a UAV at an exact point, so that it really becomes an unmanned vehicle and doesn't depend directly on the human. The project where the development is presented is entitled \Autonomous System for Localization and Action against Pollutants in the Sea (SALACOM, 2013) (DPI2013-46665-C1)" funded by the Ministry of Economy, Industry and Competitiveness where it raises the collaboration between autonomous ships and drones.The latter needing to land on a platform aboard a ship. Therefore, this project involves the recognition of a landing platform through a computer vision system based on the analysis of two types of images, thermal and color, captured by two devices, in this case installed in a Smartphone. The platform is specifically designed for this purpose. In fact, it consists of a wooden base on which five metallic figures of different colors have been placed. The geometric figures, strategically positioned between them, configure a geometric arrangement with no possibility of appearance in the operating environments of the UAV. The different colors allow geometrical elements with different spectral shapes to be discerned. Finally, its metallic nature allows different thermal reflectances in the visible spectrum to be distinguished. A color and an infrared image are always available for recognition. In this way two procedures are combined that increase the degree of accuracy in identification. This is the ideal scenario and the framework of the project, regardless of whether these same techniques are used in other contexts, not necessarily in collaboration.
Description
Trabajo de Fin de Grado en Ingeniería de Software (Universidad Complutense, Facultad de Informática, curso 2016/2017)