Aerial–Terrestrial Image Feature Matching: An Evaluation of Recent Deep Learning Methods

The 3-D reconstruction of complex urban areas is becoming increasingly important for various applications. To achieve precise and complete 3-D reconstruction, current approaches aim to combine aerial and terrestrial images. The main challenge is achieving reliable feature matching of aerial and terr...

Full description

Saved in:
Bibliographic Details
Main Authors: Hui Wang, Jiangxue Yu, San Jiang, Dejin Zhang, Qingquan Li
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10976566/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850092017933090816
author Hui Wang
Jiangxue Yu
San Jiang
Dejin Zhang
Qingquan Li
author_facet Hui Wang
Jiangxue Yu
San Jiang
Dejin Zhang
Qingquan Li
author_sort Hui Wang
collection DOAJ
description The 3-D reconstruction of complex urban areas is becoming increasingly important for various applications. To achieve precise and complete 3-D reconstruction, current approaches aim to combine aerial and terrestrial images. The main challenge is achieving reliable feature matching of aerial and terrestrial images under large viewing angles and varying scene illuminations. Traditional handcrafted methods experience a significant decline in matching performance. In this context, deep-learning-based feature matching methods have developed rapidly and gained extensive attention. However, their performance in handling challenging large-angle aerial–terrestrial datasets still needs to be evaluated. To assess their performance for aerial–terrestrial images, this study has reviewed and evaluated four types of recent deep-learning-based feature matching networks and selected four sets of aerial–terrestrial datasets for experimental tests. Extensive experiments and evaluations have been conducted in terms of feature matching and image orientation based on structure from motion (SfM). The results demonstrate that graph-neural-network-based methods and detector-free methods exhibit significant advantages in feature matching of aerial–terrestrial datasets, which can generate effective and correct matches for aerial–terrestrial images with large-scale and viewpoint differences. In particular, the combination of SuperPoint and LightGlue achieves the best performance, which can generate approximately ten times the number of aerial–terrestrial feature matches when compared with scale invariant feature transform (SIFT). In addition, all images can be registered in SfM reconstruction using its matching results. However, the precision of deep-learning-based methods is still inferior to the classical handcrafted method in SfM reconstruction. Thus, there is still significant room for improvement to enhance their performance further.
format Article
id doaj-art-1492b64310fc4d1babfe67e8a651b55a
institution DOAJ
issn 1939-1404
2151-1535
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
spelling doaj-art-1492b64310fc4d1babfe67e8a651b55a2025-08-20T02:42:12ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-0118156881570610.1109/JSTARS.2025.356432610976566Aerial–Terrestrial Image Feature Matching: An Evaluation of Recent Deep Learning MethodsHui Wang0Jiangxue Yu1San Jiang2https://orcid.org/0000-0002-7799-650XDejin Zhang3https://orcid.org/0000-0002-7423-2328Qingquan Li4https://orcid.org/0000-0002-2438-6046School of Computer Science, China University of Geosciences, Wuhan, ChinaSchool of Computer Science, China University of Geosciences, Wuhan, ChinaGuangdong Key Laboratory of Urban Informatics, Shenzhen University, Shenzhen, ChinaGuangdong Key Laboratory of Urban Informatics, Shenzhen University, Shenzhen, ChinaGuangdong Key Laboratory of Urban Informatics, Shenzhen University, Shenzhen, ChinaThe 3-D reconstruction of complex urban areas is becoming increasingly important for various applications. To achieve precise and complete 3-D reconstruction, current approaches aim to combine aerial and terrestrial images. The main challenge is achieving reliable feature matching of aerial and terrestrial images under large viewing angles and varying scene illuminations. Traditional handcrafted methods experience a significant decline in matching performance. In this context, deep-learning-based feature matching methods have developed rapidly and gained extensive attention. However, their performance in handling challenging large-angle aerial–terrestrial datasets still needs to be evaluated. To assess their performance for aerial–terrestrial images, this study has reviewed and evaluated four types of recent deep-learning-based feature matching networks and selected four sets of aerial–terrestrial datasets for experimental tests. Extensive experiments and evaluations have been conducted in terms of feature matching and image orientation based on structure from motion (SfM). The results demonstrate that graph-neural-network-based methods and detector-free methods exhibit significant advantages in feature matching of aerial–terrestrial datasets, which can generate effective and correct matches for aerial–terrestrial images with large-scale and viewpoint differences. In particular, the combination of SuperPoint and LightGlue achieves the best performance, which can generate approximately ten times the number of aerial–terrestrial feature matches when compared with scale invariant feature transform (SIFT). In addition, all images can be registered in SfM reconstruction using its matching results. However, the precision of deep-learning-based methods is still inferior to the classical handcrafted method in SfM reconstruction. Thus, there is still significant room for improvement to enhance their performance further.https://ieeexplore.ieee.org/document/10976566/Aerial–terrestrial imagesdeep learningdetector-based methodsdetector-free methodsfeature matchingstructure from motion (SfM)
spellingShingle Hui Wang
Jiangxue Yu
San Jiang
Dejin Zhang
Qingquan Li
Aerial–Terrestrial Image Feature Matching: An Evaluation of Recent Deep Learning Methods
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Aerial–terrestrial images
deep learning
detector-based methods
detector-free methods
feature matching
structure from motion (SfM)
title Aerial–Terrestrial Image Feature Matching: An Evaluation of Recent Deep Learning Methods
title_full Aerial–Terrestrial Image Feature Matching: An Evaluation of Recent Deep Learning Methods
title_fullStr Aerial–Terrestrial Image Feature Matching: An Evaluation of Recent Deep Learning Methods
title_full_unstemmed Aerial–Terrestrial Image Feature Matching: An Evaluation of Recent Deep Learning Methods
title_short Aerial–Terrestrial Image Feature Matching: An Evaluation of Recent Deep Learning Methods
title_sort aerial x2013 terrestrial image feature matching an evaluation of recent deep learning methods
topic Aerial–terrestrial images
deep learning
detector-based methods
detector-free methods
feature matching
structure from motion (SfM)
url https://ieeexplore.ieee.org/document/10976566/
work_keys_str_mv AT huiwang aerialx2013terrestrialimagefeaturematchinganevaluationofrecentdeeplearningmethods
AT jiangxueyu aerialx2013terrestrialimagefeaturematchinganevaluationofrecentdeeplearningmethods
AT sanjiang aerialx2013terrestrialimagefeaturematchinganevaluationofrecentdeeplearningmethods
AT dejinzhang aerialx2013terrestrialimagefeaturematchinganevaluationofrecentdeeplearningmethods
AT qingquanli aerialx2013terrestrialimagefeaturematchinganevaluationofrecentdeeplearningmethods