{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,9,23]],"date-time":"2025-09-23T00:14:46Z","timestamp":1758586486912,"version":"3.44.0"},"reference-count":38,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2025,9,22]],"date-time":"2025-09-22T00:00:00Z","timestamp":1758499200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["T2121003 52375003 U22B2080"],"award-info":[{"award-number":["T2121003 52375003 U22B2080"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Robot. AI"],"abstract":"<jats:sec><jats:title>Introduction<\/jats:title><jats:p>Visual odometry (VO) has been widely deployed on mobile robots for spatial perception. State-of-the-art VO offers robust localization, the maps it generates are often too sparse for downstream tasks due to insufffcient depth data. While depth completion methods can estimate dense depth from sparse data, the extreme sparsity and highly uneven distribution of depth signals in VO (\u223c 0.15% of the pixels in the depth image available) poses signiffcant challenges.<\/jats:p><\/jats:sec><jats:sec><jats:title>Methods<\/jats:title><jats:p>To address this issue, we propose a lightweight Image-Guided Uncertainty-Aware Depth Completion Network (IU-DC) for completing sparse depth from VO. This network integrates color and spatial information into a normalized convolutional neural network to tackle the sparsity issue and simultaneously outputs dense depth and associated uncertainty. The estimated depth is uncertainty-aware, allowing for the filtering of outliers and ensuring precise spatial perception.<\/jats:p><\/jats:sec><jats:sec><jats:title>Results<\/jats:title><jats:p>The superior performance of IU-DC compared to SOTA is validated across multiple open-source datasets in terms of depth and uncertainty estimation accuracy. In real-world mapping tasks, by integrating IU-DC with the mapping module, we achieve 50 \u00d7 more reconstructed volumes and 78% coverage of the ground truth with twice the accuracy compared to SOTA, despite having only 0.6 M parameters (just 3% of the size of the SOTA).<\/jats:p><\/jats:sec><jats:sec><jats:title>Discussion<\/jats:title><jats:p>Our code will be released at <jats:ext-link>https:\/\/github.com\/YangDL-BEIHANG\/Dense-mapping-from-sparse-visual-odometry\/tree\/d5a11b4403b5ac2e9e0c3644b14b9711c2748bf9<\/jats:ext-link>.<\/jats:p><\/jats:sec>","DOI":"10.3389\/frobt.2025.1644230","type":"journal-article","created":{"date-parts":[[2025,9,22]],"date-time":"2025-09-22T04:12:13Z","timestamp":1758514333000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Dense mapping from sparse visual odometry: a lightweight uncertainty-guaranteed depth completion method"],"prefix":"10.3389","volume":"12","author":[{"given":"Daolong","family":"Yang","sequence":"first","affiliation":[]},{"given":"Xudong","family":"Zhang","sequence":"additional","affiliation":[]},{"given":"Haoyuan","family":"Liu","sequence":"additional","affiliation":[]},{"given":"Haoyang","family":"Wu","sequence":"additional","affiliation":[]},{"given":"Chengcai","family":"Wang","sequence":"additional","affiliation":[]},{"given":"Kun","family":"Xu","sequence":"additional","affiliation":[]},{"given":"Xilun","family":"Ding","sequence":"additional","affiliation":[]}],"member":"1965","published-online":{"date-parts":[[2025,9,22]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"1352","DOI":"10.1177\/0278364919839762","article-title":"Confidence-rich grid mapping","volume":"38","author":"Agha-Mohammadi","year":"2019","journal-title":"Int. J. Robotics Res."},{"key":"B2","doi-asserted-by":"publisher","first-page":"832165","DOI":"10.3389\/frobt.2022.832165","article-title":"Localization and mapping on agriculture based on point-feature extraction and semiplanes segmentation from 3d lidar data","volume":"9","author":"Aguiar","year":"2022","journal-title":"Front. Robotics AI"},{"key":"B3","doi-asserted-by":"publisher","first-page":"1424036","DOI":"10.3389\/frobt.2024.1424036","article-title":"A fast monocular 6d pose estimation method for textureless objects based on perceptual hashing and template matching","volume":"11","author":"Araya-Martinez","year":"2025","journal-title":"Front. Robotics AI"},{"key":"B4","first-page":"8853","article-title":"Agg-net: attention guided gated-convolutional network for depth image completion","volume-title":"Proceedings of the IEEE\/CVF international conference on computer vision","author":"Chen","year":"2023"},{"key":"B5","doi-asserted-by":"publisher","first-page":"2423","DOI":"10.1109\/tpami.2019.2929170","article-title":"Confidence propagation through cnns for guided sparse depth regression","volume":"42","author":"Eldesokey","year":"2019","journal-title":"IEEE Trans. pattern analysis Mach. Intell."},{"key":"B6","first-page":"12 014","article-title":"Uncertainty-aware cnns for depth completion: uncertainty from beginning to end","volume-title":"Proceedings of the IEEE\/CVF conference on computer vision and pattern recognition","author":"Eldesokey","year":"2020"},{"key":"B7","doi-asserted-by":"publisher","first-page":"1661","DOI":"10.1109\/lra.2019.2896963","article-title":"Geo-supervised visual depth prediction","volume":"4","author":"Fei","year":"2019","journal-title":"IEEE Robotics Automation Lett."},{"key":"B8","first-page":"318","article-title":"Evaluating scalable bayesian deep learning methods for robust computer vision","volume-title":"Proceedings of the IEEE\/CVF conference on computer vision and pattern recognition workshops","author":"Gustafsson","year":"2020"},{"key":"B9","doi-asserted-by":"publisher","first-page":"8244","DOI":"10.1109\/TPAMI.2022.3229090","article-title":"Deep depth completion from extremely sparse data: a survey","volume":"45","author":"Hu","year":"2022","journal-title":"IEEE Trans. Pattern Analysis Mach. Intell."},{"key":"B10","first-page":"652","article-title":"Uncertainty estimates and multi-hypotheses networks for optical flow","volume-title":"Proceedings of the European conference on computer vision (ECCV)","author":"Ilg","year":"2018"},{"key":"B11","article-title":"What uncertainties do we need in bayesian deep learning for computer vision?","volume":"30","author":"Kendall","year":"2017","journal-title":"Adv. neural Inf. Process. Syst."},{"key":"B12","doi-asserted-by":"crossref","first-page":"515","DOI":"10.1109\/CVPR.1993.341081","article-title":"Normalized and differential convolution","volume-title":"Proceedings of IEEE conference on computer vision and pattern recognition","author":"Knutsson","year":"1993"},{"key":"B13","doi-asserted-by":"publisher","first-page":"416","DOI":"10.1002\/rob.21831","article-title":"Rtab-map as an open-source lidar and visual simultaneous localization and mapping library for large-scale and long-term online operation","volume":"36","author":"Labb\u00e9","year":"2019","journal-title":"J. field robotics"},{"key":"B14","doi-asserted-by":"publisher","first-page":"801886","DOI":"10.3389\/frobt.2022.801886","article-title":"Multi-session visual slam for illumination-invariant re-localization in indoor environments","volume":"9","author":"Labb\u00e9","year":"2022","journal-title":"Front. Robotics AI"},{"key":"B15","first-page":"35","article-title":"Monitored distillation for positive congruent depth completion","volume-title":"European conference on computer vision","author":"Liu","year":"2022"},{"key":"B16","doi-asserted-by":"publisher","first-page":"920","DOI":"10.1109\/lra.2023.3234776","article-title":"Mff-net: towards efficient monocular depth completion with multi-modal feature fusion","volume":"8","author":"Liu","year":"2023","journal-title":"IEEE Robotics Automation Lett."},{"key":"B17","doi-asserted-by":"crossref","first-page":"380","DOI":"10.1109\/3DV.2016.47","article-title":"A closed-form bayesian fusion equation using occupancy probabilities","volume-title":"2016 fourth international conference on 3D vision (3DV)","author":"Loop","year":"2016"},{"key":"B18","doi-asserted-by":"crossref","first-page":"4796","DOI":"10.1109\/ICRA.2018.8460184","article-title":"Sparse-to-dense: depth prediction from sparse depth samples and a single image","volume-title":"2018 IEEE international conference on robotics and automation (ICRA)","author":"Ma","year":"2018"},{"key":"B19","doi-asserted-by":"publisher","first-page":"1291672","DOI":"10.3389\/frobt.2023.1291672","article-title":"Simultaneous localization and mapping in a multi-robot system in a dynamic environment with unknown initial correspondence","volume":"10","author":"Malakouti-Khah","year":"2024","journal-title":"Front. Robotics AI"},{"key":"B20","doi-asserted-by":"publisher","first-page":"1212525","DOI":"10.3389\/frobt.2023.1212525","article-title":"Self-supervised monocular depth estimation for high field of view colonoscopy cameras","volume":"10","author":"Mathew","year":"2023","journal-title":"Front. Robotics AI"},{"key":"B21","doi-asserted-by":"crossref","first-page":"5713","DOI":"10.1109\/ICRA48506.2021.9561174","article-title":"Robust monocular visual-inertial depth completion for embedded systems","volume-title":"2021 IEEE international conference on robotics and automation (ICRA)","author":"Merrill","year":"2021"},{"key":"B22","doi-asserted-by":"publisher","first-page":"5072","DOI":"10.1109\/lra.2021.3070308","article-title":"Volumetric occupancy mapping with probabilistic depth completion for robotic navigation","volume":"6","author":"Popovi\u0107","year":"2021","journal-title":"IEEE Robotics Automation Lett."},{"key":"B23","doi-asserted-by":"publisher","first-page":"1004","DOI":"10.1109\/tro.2018.2853729","article-title":"Vins-mono: a robust and versatile monocular visual-inertial state estimator","volume":"34","author":"Qin","year":"2018","journal-title":"IEEE Trans. Robotics"},{"key":"B24","first-page":"16 147","article-title":"Bayesian deep basis fitting for depth completion with uncertainty","volume-title":"Proceedings of the IEEE\/CVF international conference on computer vision","author":"Qu","year":"2021"},{"key":"B25","first-page":"234","article-title":"U-net: convolutional networks for biomedical image segmentation","volume-title":"Medical image computing and computer-assisted intervention\u2013MICCAI 2015: 18Th international conference, munich, Germany, October 5-9, 2015, proceedings, part III 18","author":"Ronneberger","year":"2015"},{"key":"B26","volume-title":"The Robotics: Science and Systems (RSS)","author":"Segal","year":"2009"},{"key":"B27","doi-asserted-by":"crossref","first-page":"746","DOI":"10.1007\/978-3-642-33715-4_54","article-title":"Indoor segmentation and support inference from rgbd images","volume-title":"Computer Vision\u2013ECCV 2012: 12Th European conference on computer vision, florence, Italy, October 7-13, 2012, proceedings, part V 12","author":"Silberman","year":"2012"},{"key":"B28","first-page":"3483","article-title":"Frame: fast and robust autonomous 3d point cloud map-merging for egocentric multi-robot exploration","author":"Stathoulopoulos","year":"2023"},{"key":"B29","doi-asserted-by":"crossref","first-page":"5339","DOI":"10.1109\/IROS47612.2022.9981531","article-title":"3d lidar reconstruction with probabilistic depth completion for robotic navigation","volume-title":"2022 IEEE\/RSJ international conference on intelligent robots and systems (IROS)","author":"Tao","year":"2022"},{"key":"B30","first-page":"16 558","article-title":"Droid-slam: deep visual slam for monocular, stereo, and rgb-d cameras","volume":"34","author":"Teed","year":"2021","journal-title":"Adv. neural Inf. Process. Syst."},{"key":"B31","doi-asserted-by":"publisher","first-page":"1055","DOI":"10.1109\/lra.2020.2967296","article-title":"Aerial single-view depth completion with image-guided uncertainty estimation","volume":"5","author":"Teixeira","year":"2020","journal-title":"IEEE Robotics Automation Lett."},{"key":"B32","doi-asserted-by":"crossref","DOI":"10.1109\/3DV.2017.00012","article-title":"Sparsity invariant cnns","volume-title":"International conference on 3D vision (3DV)","author":"Uhrig","year":"2017"},{"key":"B33","doi-asserted-by":"publisher","first-page":"5684","DOI":"10.1109\/tcsvt.2025.3527235","article-title":"Mixssc: forward-backward mixture for vision-based 3d semantic scene completion","volume":"35","author":"Wang","year":"2025","journal-title":"IEEE Trans. Circuits Syst. Video Technol."},{"key":"B34","doi-asserted-by":"crossref","first-page":"6095","DOI":"10.1109\/ICRA48891.2023.10161013","article-title":"Monocular visual-inertial depth estimation","volume-title":"2023 IEEE international conference on robotics and automation (ICRA)","author":"Wofk","year":"2023"},{"key":"B35","doi-asserted-by":"publisher","first-page":"1899","DOI":"10.1109\/lra.2020.2969938","article-title":"Unsupervised depth completion from visual inertial odometry","volume":"5","author":"Wong","year":"2020","journal-title":"IEEE Robotics Automation Lett."},{"key":"B36","first-page":"3","article-title":"Cbam: convolutional block attention module","volume-title":"Proceedings of the European conference on computer vision (ECCV)","author":"Woo","year":"2018"},{"key":"B37","first-page":"4471","article-title":"Free-form image inpainting with gated convolution","volume-title":"Proceedings of the IEEE\/CVF international conference on computer vision","author":"Yu","year":"2019"},{"key":"B38","first-page":"5002","article-title":"Dui-vio: depth uncertainty incorporated visual inertial odometry based on an rgb-d camera","author":"Zhang","year":"2020"}],"container-title":["Frontiers in Robotics and AI"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frobt.2025.1644230\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,9,22]],"date-time":"2025-09-22T04:12:18Z","timestamp":1758514338000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frobt.2025.1644230\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,9,22]]},"references-count":38,"alternative-id":["10.3389\/frobt.2025.1644230"],"URL":"https:\/\/doi.org\/10.3389\/frobt.2025.1644230","relation":{},"ISSN":["2296-9144"],"issn-type":[{"value":"2296-9144","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,9,22]]},"article-number":"1644230"}}