HFNet-SLAM: An Accurate and Real-Time Monocular SLAM System with Deep Features
Image tracking and retrieval strategies are of vital importance in visual Simultaneous Localization and Mapping (SLAM) systems. For most state-of-the-art systems, hand-crafted features and bag-of-words (BoW) algorithms are the common solutions. Recent research reports the vulnerability of these trad...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-02-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/23/4/2113 |
| _version_ | 1797618212411539456 |
|---|---|
| author | Liming Liu Jonathan M. Aitken |
| author_facet | Liming Liu Jonathan M. Aitken |
| author_sort | Liming Liu |
| collection | DOAJ |
| description | Image tracking and retrieval strategies are of vital importance in visual Simultaneous Localization and Mapping (SLAM) systems. For most state-of-the-art systems, hand-crafted features and bag-of-words (BoW) algorithms are the common solutions. Recent research reports the vulnerability of these traditional algorithms in complex environments. To replace these methods, this work proposes HFNet-SLAM, an accurate and real-time monocular SLAM system built on the ORB-SLAM3 framework incorporated with deep convolutional neural networks (CNNs). This work provides a pipeline of feature extraction, keypoint matching, and loop detection fully based on features from CNNs. The performance of this system has been validated on public datasets against other state-of-the-art algorithms. The results reveal that the HFNet-SLAM achieves the lowest errors among systems available in the literature. Notably, the HFNet-SLAM obtains an average accuracy of 2.8 cm in EuRoC dataset in pure visual configuration. Besides, it doubles the accuracy in medium and large environments in TUM-VI dataset compared with ORB-SLAM3. Furthermore, with the optimisation of TensorRT technology, the entire system can run in real-time at 50 FPS. |
| first_indexed | 2024-03-11T08:10:58Z |
| format | Article |
| id | doaj.art-853b8ad8787647978e4e555e317a5bd4 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-11T08:10:58Z |
| publishDate | 2023-02-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-853b8ad8787647978e4e555e317a5bd42023-11-16T23:10:27ZengMDPI AGSensors1424-82202023-02-01234211310.3390/s23042113HFNet-SLAM: An Accurate and Real-Time Monocular SLAM System with Deep FeaturesLiming Liu0Jonathan M. Aitken1Department of Automatic Control and Systems Engineering, The University of Sheffield, Sheffield S10 2TN, UKDepartment of Automatic Control and Systems Engineering, The University of Sheffield, Sheffield S10 2TN, UKImage tracking and retrieval strategies are of vital importance in visual Simultaneous Localization and Mapping (SLAM) systems. For most state-of-the-art systems, hand-crafted features and bag-of-words (BoW) algorithms are the common solutions. Recent research reports the vulnerability of these traditional algorithms in complex environments. To replace these methods, this work proposes HFNet-SLAM, an accurate and real-time monocular SLAM system built on the ORB-SLAM3 framework incorporated with deep convolutional neural networks (CNNs). This work provides a pipeline of feature extraction, keypoint matching, and loop detection fully based on features from CNNs. The performance of this system has been validated on public datasets against other state-of-the-art algorithms. The results reveal that the HFNet-SLAM achieves the lowest errors among systems available in the literature. Notably, the HFNet-SLAM obtains an average accuracy of 2.8 cm in EuRoC dataset in pure visual configuration. Besides, it doubles the accuracy in medium and large environments in TUM-VI dataset compared with ORB-SLAM3. Furthermore, with the optimisation of TensorRT technology, the entire system can run in real-time at 50 FPS.https://www.mdpi.com/1424-8220/23/4/2113simultaneous location and mappingdeep featuresmonocular localisation |
| spellingShingle | Liming Liu Jonathan M. Aitken HFNet-SLAM: An Accurate and Real-Time Monocular SLAM System with Deep Features Sensors simultaneous location and mapping deep features monocular localisation |
| title | HFNet-SLAM: An Accurate and Real-Time Monocular SLAM System with Deep Features |
| title_full | HFNet-SLAM: An Accurate and Real-Time Monocular SLAM System with Deep Features |
| title_fullStr | HFNet-SLAM: An Accurate and Real-Time Monocular SLAM System with Deep Features |
| title_full_unstemmed | HFNet-SLAM: An Accurate and Real-Time Monocular SLAM System with Deep Features |
| title_short | HFNet-SLAM: An Accurate and Real-Time Monocular SLAM System with Deep Features |
| title_sort | hfnet slam an accurate and real time monocular slam system with deep features |
| topic | simultaneous location and mapping deep features monocular localisation |
| url | https://www.mdpi.com/1424-8220/23/4/2113 |
| work_keys_str_mv | AT limingliu hfnetslamanaccurateandrealtimemonocularslamsystemwithdeepfeatures AT jonathanmaitken hfnetslamanaccurateandrealtimemonocularslamsystemwithdeepfeatures |