A deep learning-based pipeline for visual geolocation in the urban environment

dc.contributor.author	Tsapiv, Volodymyr
dc.date.accessioned	2024-02-14T11:37:42Z
dc.date.available	2024-02-14T11:37:42Z
dc.date.issued	2023
dc.identifier.citation	Tsapiv, Volodymyr. A deep learning-based pipeline for visual geolocation in the urban environment / Volodymyr Tsapiv; Supervisor: Taras Firman; Ukrainian Catholic University, Department of Computer Sciences. – Lviv: 2023. – 42 p.: ill.	uk
dc.identifier.uri	https://er.ucu.edu.ua/handle/1/4428
dc.language.iso	en	uk
dc.title	A deep learning-based pipeline for visual geolocation in the urban environment	uk
dc.type	Preprint	uk
dc.status	Публікується вперше	uk
dc.description.abstracten	The precise identification of location in urban areas is a challenging problem for Global Navigation Satellite Systems (GNSS), such as GPS, because of obstacles that include signal blockage, multipath interference, and urban canyons, among other factors. This thesis proposes a structure-based visual localization pipeline that uses a combination of Deep Neural Networks (DNNs) and traditional computer vision algorithms to perform accurate localization by an image. Additionally, we provide a collection of helpful tools for constructing a reference database for visual localization that can be used with any city found on Google Maps. The proposed method was evaluated on established visual localization benchmarks and produced competitive outcomes.	uk