ARVC ES    EN
Photo

Vicente Román

Researcher
Tel:  +34 96 522 2435
Email: vicente.roman@goumh.umh.es

My Projects

Research Projects

EMERG2020

Title: RECONSTRUCTION OF SCENES FROM OMNIDIRECTIONAL CAMERAS USING VISUAL APPEARANCE AND DEEP LEARNING TECHNIQUES.

Funded by: Generalitat Valenciana

Duration: 01/2020 - 12/2020

Description:


Most of the existing algorithms that solve mapping and location problems stop working properly when the robot operates in an unstructured, complex and changing environment or when the robot can move with more than three degrees of freedom (DOF). ). In response to this challenge, the main research line of this project proposes the improvement and development of new mechanisms that allow efficient, robust and precise modeling of environments using vision systems. Specifically, the use of omnidirectional vision systems is proposed due to the large amount of information they provide at a relatively low cost. However, the use of these vision systems makes it necessary to consider the challenges of working with the images provided by this type of camera. In this sense, it is proposed to study in depth descriptors based on global appearance and make use of Deep Learning techniques.
The development of this project is developed through various objectives such as the analysis of the present algorithms for creating maps and location, comparison of the present global appearance algorithms and also, developing new location algorithms and / or appearance descriptors global based on Deep Learning. In order to improve the integration of the mobile robot in real work environments (Industry 4.0), in which they interact with people, characteristics that make it compatible with human perception will be incorporated into the map.

Keywords: Deep learning, scene reconstruction, localization, omnidirectional vision

Head Researcher: M. Ballesta

More Information...     




© ARVC - UMH

Miguel Hernández University