A Review on Deep Learning for UAV Absolute Visual Localization

In the past few years, the use of Unmanned Aerial Vehicles (UAVs) has expanded and now reached mainstream levels for applications such as infrastructure inspection, agriculture, transport, security, entertainment, real estate, environmental conservation, search and rescue, and even insurance. This s...

Full description

Saved in:
Bibliographic Details
Main Authors: Andy Couturier, Moulay A. Akhloufi
Format: Article
Language:English
Published: MDPI AG 2024-10-01
Series:Drones
Subjects:
Online Access:https://www.mdpi.com/2504-446X/8/11/622
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850217421152976896
author Andy Couturier
Moulay A. Akhloufi
author_facet Andy Couturier
Moulay A. Akhloufi
author_sort Andy Couturier
collection DOAJ
description In the past few years, the use of Unmanned Aerial Vehicles (UAVs) has expanded and now reached mainstream levels for applications such as infrastructure inspection, agriculture, transport, security, entertainment, real estate, environmental conservation, search and rescue, and even insurance. This surge in adoption can be attributed to the UAV ecosystem’s maturation, which has not only made these devices more accessible and cost effective but has also significantly enhanced their operational capabilities in terms of flight duration and embedded computing power. In conjunction with these developments, the research on Absolute Visual Localization (AVL) has seen a resurgence driven by the introduction of deep learning to the field. These new approaches have significantly improved localization solutions in comparison to the previous generation of approaches based on traditional computer vision feature extractors. This paper conducts an extensive review of the literature on deep learning-based methods for UAV AVL, covering significant advancements since 2019. It retraces key developments that have led to the rise in learning-based approaches and provides an in-depth analysis of related localization sources such as Inertial Measurement Units (IMUs) and Global Navigation Satellite Systems (GNSSs), highlighting their limitations and advantages for more effective integration with AVL. The paper concludes with an analysis of current challenges and proposes future research directions to guide further work in the field.
format Article
id doaj-art-bb0c81aae6b749f4b1ba2ce12255bd87
institution OA Journals
issn 2504-446X
language English
publishDate 2024-10-01
publisher MDPI AG
record_format Article
series Drones
spelling doaj-art-bb0c81aae6b749f4b1ba2ce12255bd872025-08-20T02:08:03ZengMDPI AGDrones2504-446X2024-10-0181162210.3390/drones8110622A Review on Deep Learning for UAV Absolute Visual LocalizationAndy Couturier0Moulay A. Akhloufi1Computer Science Department, Science Faculty, Université de Moncton, 18 Antonine-Maillet Ave, Moncton, NB E1A 3E9, CanadaComputer Science Department, Science Faculty, Université de Moncton, 18 Antonine-Maillet Ave, Moncton, NB E1A 3E9, CanadaIn the past few years, the use of Unmanned Aerial Vehicles (UAVs) has expanded and now reached mainstream levels for applications such as infrastructure inspection, agriculture, transport, security, entertainment, real estate, environmental conservation, search and rescue, and even insurance. This surge in adoption can be attributed to the UAV ecosystem’s maturation, which has not only made these devices more accessible and cost effective but has also significantly enhanced their operational capabilities in terms of flight duration and embedded computing power. In conjunction with these developments, the research on Absolute Visual Localization (AVL) has seen a resurgence driven by the introduction of deep learning to the field. These new approaches have significantly improved localization solutions in comparison to the previous generation of approaches based on traditional computer vision feature extractors. This paper conducts an extensive review of the literature on deep learning-based methods for UAV AVL, covering significant advancements since 2019. It retraces key developments that have led to the rise in learning-based approaches and provides an in-depth analysis of related localization sources such as Inertial Measurement Units (IMUs) and Global Navigation Satellite Systems (GNSSs), highlighting their limitations and advantages for more effective integration with AVL. The paper concludes with an analysis of current challenges and proposes future research directions to guide further work in the field.https://www.mdpi.com/2504-446X/8/11/622unmanned aerial vehiclesdeep learninglocalizationGNSS-denied navigation
spellingShingle Andy Couturier
Moulay A. Akhloufi
A Review on Deep Learning for UAV Absolute Visual Localization
Drones
unmanned aerial vehicles
deep learning
localization
GNSS-denied navigation
title A Review on Deep Learning for UAV Absolute Visual Localization
title_full A Review on Deep Learning for UAV Absolute Visual Localization
title_fullStr A Review on Deep Learning for UAV Absolute Visual Localization
title_full_unstemmed A Review on Deep Learning for UAV Absolute Visual Localization
title_short A Review on Deep Learning for UAV Absolute Visual Localization
title_sort review on deep learning for uav absolute visual localization
topic unmanned aerial vehicles
deep learning
localization
GNSS-denied navigation
url https://www.mdpi.com/2504-446X/8/11/622
work_keys_str_mv AT andycouturier areviewondeeplearningforuavabsolutevisuallocalization
AT moulayaakhloufi areviewondeeplearningforuavabsolutevisuallocalization
AT andycouturier reviewondeeplearningforuavabsolutevisuallocalization
AT moulayaakhloufi reviewondeeplearningforuavabsolutevisuallocalization