State-of-the-art in Visual Geo-localization

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216305%3A26230%2F17%3APU123185" target="_blank" >RIV/00216305:26230/17:PU123185 - isvavai.cz</a>

Výsledek na webu

<a href="http://cadik.posvete.cz/papers/brejcha-cadik17geolocalization_methods_survey.pdf" target="_blank" >http://cadik.posvete.cz/papers/brejcha-cadik17geolocalization_methods_survey.pdf</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1007/s10044-017-0611-1" target="_blank" >10.1007/s10044-017-0611-1</a>

Jazyk výsledku

angličtina

Název v původním jazyce

State-of-the-art in Visual Geo-localization

Popis výsledku v původním jazyce

Large-scale visual geo-localization has recently gained a lot of attention in computer vision research and new methods are proposed steadily. However, surveys of visual geo-localization methods are rare and they focus mainly on city-scale localization methods. We present a comprehensive and balanced study of existing visual geo-localization domains, including city-scale, global approaches and methods for natural environments. We overview the methods to show their pros and cons, application domains, datasets, as well as evaluation techniques. We categorize the reviewed methods by two criteria. The first is the type of data the method uses for geo-location estimation. The second criterion is the target environment for which the method has been proposed and validated. Based on this categorization we analyze important conditions that must be considered while solving geo-localization problems. Each category is in a different state of research - while city-scale image-based methods received a lot of attention, other categories like natural environments using cross-domain data sources are still challenging problems under active research. Future research of large-scale visual geo-localization is discussed, primarily the challenging and new research category - geo-localization in natural environments.

Název v anglickém jazyce

State-of-the-art in Visual Geo-localization

Popis výsledku anglicky

Large-scale visual geo-localization has recently gained a lot of attention in computer vision research and new methods are proposed steadily. However, surveys of visual geo-localization methods are rare and they focus mainly on city-scale localization methods. We present a comprehensive and balanced study of existing visual geo-localization domains, including city-scale, global approaches and methods for natural environments. We overview the methods to show their pros and cons, application domains, datasets, as well as evaluation techniques. We categorize the reviewed methods by two criteria. The first is the type of data the method uses for geo-location estimation. The second criterion is the target environment for which the method has been proposed and validated. Based on this categorization we analyze important conditions that must be considered while solving geo-localization problems. Each category is in a different state of research - while city-scale image-based methods received a lot of attention, other categories like natural environments using cross-domain data sources are still challenging problems under active research. Future research of large-scale visual geo-localization is discussed, primarily the challenging and new research category - geo-localization in natural environments.

Druh

J_imp - Článek v periodiku v databázi Web of Science

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2017

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název periodika

PATTERN ANALYSIS AND APPLICATIONS

ISSN

1433-7541

e-ISSN

1433-755X

Svazek periodika

2017

Číslo periodika v rámci svazku

3

Stát vydavatele periodika

GB - Spojené království Velké Británie a Severního Irska

Počet stran výsledku

25

Strana od-do

1-25

Kód UT WoS článku

000405607000001

EID výsledku v databázi Scopus

2-s2.0-85016120066