Use this link to cite:
http://hdl.handle.net/2183/23709 Evaluación de nuevos modos de empleo de los descriptores de apariencia global en tareas de localización
Loading...
Identifiers
Publication date
Authors
Román, Vicente
Cebollada, Sergio
Payá, Luis
Tenza, María Flores
Gil, Arturo
Reinoso, Óscar
Advisors
Other responsabilities
Journal Title
Bibliographic citation
Román Erades, V., Sergio Cebollada Lopez, S., Paya, L., Tenza, M.F., Gil Aparicio, A., Reinoso, O. (2019). Evaluación de nuevos modos de empleo de los descriptores de apariencia global en tareas de localización. En XL Jornadas de Automática: libro de actas, Ferrol, 4-6 de septiembre de 2019 (pp. 842-848). DOI capítulo: https://doi.org/10.17979/spudc.9788497497169.842. DOI libro: https://doi.org/10.17979/spudc.9788497497169
Type of academic work
Academic degree
Abstract
[Resumen] Los robots aut onomos deben ser competentes en
tareas de localizaci on y creaci on de mapas. Una de
las posibles soluciones para estos problemas es em-
plear descriptores de apariencia global que obtie-
nen un unico vector que describe globalmente una
imagen panor amica previamente adquirida. Com-
parando los descriptores HOG y Gist, el prop osito
de este trabajo es el estudio de un nuevo modo
de utilizar estos descriptores con la nalidad de
sustituir el modo de empleo actual o lo que ser a
mas interesante combinar ambos modos para ob-
tener un m etodo de mayor calidad. Este trabajo se
realiza con im agenes reales tomadas en un cam-
po de trabajo heterog eneo en el que simult anea-
mente conviven personas y robots, por ello en las
im agenes hay cambios de iluminaci on, de mobilia-
rio y personas que ocluyen parte de la escena. Co-
mo conclusiones se obtiene que el descriptor HOG
tiene un resultado similar empleando ambos mo-
dos mientras que con el descriptor Gist se obtiene
un peor resultado al utilizar el nuevo modo de ven-
tanas verticales.
[Abstract] Map building and localization are two im- portant tasks that autonomous mobile ro- bots have to deal with. In the last deca- des it has appeared many studies and tech- niques to approach these problems. This work brie y describes some possible so- lutions and it recommends global appea- rance descriptors to carry our localization tasks. The aim is to obtain a unique vec- tor that describes globally the panoramic image. The mobile robot is able to estimate its position and to create maps using these methods, the result will be independent of the robot orientation. This paper goal is to use a brand new way to use the global ap- pearance descriptors. The nal result could be to use this new way or, a more interes- ting result, use a combination between both ways taken as a result a third better way. This work will consist on a comparison bet- ween both global appearance method ways. The study has been carried out using real images taken in an heterogeneous atmosp- here where humans and robots work toget- her, for that reason light condition changes and humans can apppear on the scene.
[Abstract] Map building and localization are two im- portant tasks that autonomous mobile ro- bots have to deal with. In the last deca- des it has appeared many studies and tech- niques to approach these problems. This work brie y describes some possible so- lutions and it recommends global appea- rance descriptors to carry our localization tasks. The aim is to obtain a unique vec- tor that describes globally the panoramic image. The mobile robot is able to estimate its position and to create maps using these methods, the result will be independent of the robot orientation. This paper goal is to use a brand new way to use the global ap- pearance descriptors. The nal result could be to use this new way or, a more interes- ting result, use a combination between both ways taken as a result a third better way. This work will consist on a comparison bet- ween both global appearance method ways. The study has been carried out using real images taken in an heterogeneous atmosp- here where humans and robots work toget- her, for that reason light condition changes and humans can apppear on the scene.
Description
Editor version
Rights
Atribución-NoComercial-CompartirIgual 4.0


