Use this link to cite:
http://hdl.handle.net/2183/31924 Análisis automático de flujo de vídeo para la detección de eventos de atraque y desatraque en entornos portuarios
Loading...
Identifiers
Publication date
Authors
Martínez Villar, Diego
Other responsabilities
Universidade da Coruña. Facultade de Informática
Journal Title
Bibliographic citation
Type of academic work
Academic degree
Abstract
[Resumen]: Los puertos marítimos se han convertido en un pilar fundamental para el desarrollo económico
global. Los efectos de la globalización han hecho que los puertos se posicionen como
el principal método de transporte de bienes materiales, lo que ha provocado que haya un
flujo comercial constante y a gran escala. Por ello, es necesario transformar estos entornos
portuarios en los llamados “Smart Ports” o puertos inteligentes para mejorar la seguridad, la
eficiencia y la supervisión de las operaciones que se llevan a cabo en ellos. En este proyecto, se
proponen dos metodologías automáticas para la detección de eventos de atraque y desatraque
de barcos, estando la primera basada en técnicas clásicas de visión artificial y la segunda en
técnicas de Deep Learning. En base a los resultados obtenidos, se analiza qué aproximación
es más adecuada para este dominio de aplicación. Finalmente, se integra este sistema en un
servicio vinculado a una plataforma ”Smart Ports” para el manejo automatizado de eventos, lo
que muestra una aplicación del sistema en el mundo real. Son escasos, si no nulos, los antecedentes
bibliográficos referentes a la problemática propuesta, lo que añade valor al desarrollo
de este sistema, dando pie a futuros avances en la investigación del dominio.
[Abstract]: Seaports have become fundamental pillars of worldwide economic development. The effects of globalization have made ports position themselves as the main method of goods transportation, which has produced a constant and large-scale commercial flow. Thus, it is necessary to transform these port environments into the so-called “Smart Ports” to improve security, efficiency and supervision involving the operations performed in said environments. In this project, two automatic methodologies for the detection of ship docking and undocking events are proposed, the first based on classical computer vision techniques and the second on Deep Learning ones. From the obtained results, which methodology between the former and the latter is more adequate for this application domain is discussed. Finally, the system is integrated into a service linked to a “Smart Ports” platform for automated event handling, showing a real-world implementation of the system. Bibliographic precedents referring to the proposed problem are scarce, if not null, which adds value to the development of this system, giving cause for future advances in the domain’s research.
[Abstract]: Seaports have become fundamental pillars of worldwide economic development. The effects of globalization have made ports position themselves as the main method of goods transportation, which has produced a constant and large-scale commercial flow. Thus, it is necessary to transform these port environments into the so-called “Smart Ports” to improve security, efficiency and supervision involving the operations performed in said environments. In this project, two automatic methodologies for the detection of ship docking and undocking events are proposed, the first based on classical computer vision techniques and the second on Deep Learning ones. From the obtained results, which methodology between the former and the latter is more adequate for this application domain is discussed. Finally, the system is integrated into a service linked to a “Smart Ports” platform for automated event handling, showing a real-world implementation of the system. Bibliographic precedents referring to the proposed problem are scarce, if not null, which adds value to the development of this system, giving cause for future advances in the domain’s research.
Description
Editor version
Rights
Atribución-NoComercial-SinDerivadas 3.0 España








