Integrating Sparse Learning-Based Feature Detectors into Simultaneous Localization and Mapping—A Benchmark Study
Simultaneous localization and mapping (SLAM) is one of the cornerstones of autonomous navigation systems in robotics and the automotive industry. Visual SLAM (V-SLAM), which relies on image features, such as keypoints and descriptors to estimate the pose transformation between consecutive frames, is...
Main Authors: | , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2023-02-01
|
Series: | Sensors |
Subjects: | |
Online Access: | https://www.mdpi.com/1424-8220/23/4/2286 |
_version_ | 1797618181891686400 |
---|---|
author | Giuseppe Mollica Marco Legittimo Alberto Dionigi Gabriele Costante Paolo Valigi |
author_facet | Giuseppe Mollica Marco Legittimo Alberto Dionigi Gabriele Costante Paolo Valigi |
author_sort | Giuseppe Mollica |
collection | DOAJ |
description | Simultaneous localization and mapping (SLAM) is one of the cornerstones of autonomous navigation systems in robotics and the automotive industry. Visual SLAM (V-SLAM), which relies on image features, such as keypoints and descriptors to estimate the pose transformation between consecutive frames, is a highly efficient and effective approach for gathering environmental information. With the rise of representation learning, feature detectors based on deep neural networks (DNNs) have emerged as an alternative to handcrafted solutions. This work examines the integration of sparse learned features into a state-of-the-art SLAM framework and benchmarks handcrafted and learning-based approaches by comparing the two methods through in-depth experiments. Specifically, we replace the ORB detector and BRIEF descriptor of the ORBSLAM3 pipeline with those provided by Superpoint, a DNN model that jointly computes keypoints and descriptors. Experiments on three publicly available datasets from different application domains were conducted to evaluate the pose estimation performance and resource usage of both solutions. |
first_indexed | 2024-03-11T08:09:31Z |
format | Article |
id | doaj.art-0d2ed87686e3474fa1fe12b8764568d1 |
institution | Directory Open Access Journal |
issn | 1424-8220 |
language | English |
last_indexed | 2024-03-11T08:09:31Z |
publishDate | 2023-02-01 |
publisher | MDPI AG |
record_format | Article |
series | Sensors |
spelling | doaj.art-0d2ed87686e3474fa1fe12b8764568d12023-11-16T23:12:48ZengMDPI AGSensors1424-82202023-02-01234228610.3390/s23042286Integrating Sparse Learning-Based Feature Detectors into Simultaneous Localization and Mapping—A Benchmark StudyGiuseppe Mollica0Marco Legittimo1Alberto Dionigi2Gabriele Costante3Paolo Valigi4Dipartimento di Ingegneria, Università degli Studi di Perugia, 06125 Perugia, ItalyDipartimento di Ingegneria, Università degli Studi di Perugia, 06125 Perugia, ItalyDipartimento di Ingegneria, Università degli Studi di Perugia, 06125 Perugia, ItalyDipartimento di Ingegneria, Università degli Studi di Perugia, 06125 Perugia, ItalyDipartimento di Ingegneria, Università degli Studi di Perugia, 06125 Perugia, ItalySimultaneous localization and mapping (SLAM) is one of the cornerstones of autonomous navigation systems in robotics and the automotive industry. Visual SLAM (V-SLAM), which relies on image features, such as keypoints and descriptors to estimate the pose transformation between consecutive frames, is a highly efficient and effective approach for gathering environmental information. With the rise of representation learning, feature detectors based on deep neural networks (DNNs) have emerged as an alternative to handcrafted solutions. This work examines the integration of sparse learned features into a state-of-the-art SLAM framework and benchmarks handcrafted and learning-based approaches by comparing the two methods through in-depth experiments. Specifically, we replace the ORB detector and BRIEF descriptor of the ORBSLAM3 pipeline with those provided by Superpoint, a DNN model that jointly computes keypoints and descriptors. Experiments on three publicly available datasets from different application domains were conducted to evaluate the pose estimation performance and resource usage of both solutions.https://www.mdpi.com/1424-8220/23/4/2286learning-based features detectorssimultaneous localization and mapping (SLAM)vision-based pose estimationdeep learning |
spellingShingle | Giuseppe Mollica Marco Legittimo Alberto Dionigi Gabriele Costante Paolo Valigi Integrating Sparse Learning-Based Feature Detectors into Simultaneous Localization and Mapping—A Benchmark Study Sensors learning-based features detectors simultaneous localization and mapping (SLAM) vision-based pose estimation deep learning |
title | Integrating Sparse Learning-Based Feature Detectors into Simultaneous Localization and Mapping—A Benchmark Study |
title_full | Integrating Sparse Learning-Based Feature Detectors into Simultaneous Localization and Mapping—A Benchmark Study |
title_fullStr | Integrating Sparse Learning-Based Feature Detectors into Simultaneous Localization and Mapping—A Benchmark Study |
title_full_unstemmed | Integrating Sparse Learning-Based Feature Detectors into Simultaneous Localization and Mapping—A Benchmark Study |
title_short | Integrating Sparse Learning-Based Feature Detectors into Simultaneous Localization and Mapping—A Benchmark Study |
title_sort | integrating sparse learning based feature detectors into simultaneous localization and mapping a benchmark study |
topic | learning-based features detectors simultaneous localization and mapping (SLAM) vision-based pose estimation deep learning |
url | https://www.mdpi.com/1424-8220/23/4/2286 |
work_keys_str_mv | AT giuseppemollica integratingsparselearningbasedfeaturedetectorsintosimultaneouslocalizationandmappingabenchmarkstudy AT marcolegittimo integratingsparselearningbasedfeaturedetectorsintosimultaneouslocalizationandmappingabenchmarkstudy AT albertodionigi integratingsparselearningbasedfeaturedetectorsintosimultaneouslocalizationandmappingabenchmarkstudy AT gabrielecostante integratingsparselearningbasedfeaturedetectorsintosimultaneouslocalizationandmappingabenchmarkstudy AT paolovaligi integratingsparselearningbasedfeaturedetectorsintosimultaneouslocalizationandmappingabenchmarkstudy |