Convolutional Neural Network-Based Robot Navigation Using Uncalibrated Spherical Images. Ran, L., Zhang, Y., Zhang, Q., & Yang, T. Sensors, 17(6):1341, June, 2017. ![Convolutional Neural Network-Based Robot Navigation Using Uncalibrated Spherical Images [link]](https://bibbase.org/img/filetypes/link.svg "link")
Paper doi abstract bibtex Vision-based mobile robot navigation is a vibrant area of research with numerous algorithms having been developed, the vast majority of which either belong to the scene-oriented simultaneous localization and mapping (SLAM) or fall into the category of robot-oriented lane-detection/trajectory tracking. These methods suffer from high computational cost and require stringent labelling and calibration efforts. To address these challenges, this paper proposes a lightweight robot navigation framework based purely on uncalibrated spherical images. To simplify the orientation estimation, path prediction and improve computational efficiency, the navigation problem is decomposed into a series of classification tasks. To mitigate the adverse effects of insufficient negative samples in the “navigation via classification” task, we introduce the spherical camera for scene capturing, which enables 360◦ fisheye panorama as training samples and generation of sufficient positive and negative heading directions. The classification is implemented as an end-to-end Convolutional Neural Network (CNN), trained on our proposed Spherical-Navi image dataset, whose category labels can be efficiently collected. This CNN is capable of predicting potential path directions with high confidence levels based on a single, uncalibrated spherical image. Experimental results demonstrate that the proposed framework outperforms competing ones in realistic applications.
@article{ran_convolutional_2017,
title = {Convolutional {Neural} {Network}-{Based} {Robot} {Navigation} {Using} {Uncalibrated} {Spherical} {Images}},
volume = {17},
issn = {1424-8220},
url = {http://www.mdpi.com/1424-8220/17/6/1341},
doi = {10.3390/s17061341},
abstract = {Vision-based mobile robot navigation is a vibrant area of research with numerous algorithms having been developed, the vast majority of which either belong to the scene-oriented simultaneous localization and mapping (SLAM) or fall into the category of robot-oriented lane-detection/trajectory tracking. These methods suffer from high computational cost and require stringent labelling and calibration efforts. To address these challenges, this paper proposes a lightweight robot navigation framework based purely on uncalibrated spherical images. To simplify the orientation estimation, path prediction and improve computational efficiency, the navigation problem is decomposed into a series of classification tasks. To mitigate the adverse effects of insufficient negative samples in the “navigation via classification” task, we introduce the spherical camera for scene capturing, which enables 360◦ fisheye panorama as training samples and generation of sufficient positive and negative heading directions. The classification is implemented as an end-to-end Convolutional Neural Network (CNN), trained on our proposed Spherical-Navi image dataset, whose category labels can be efficiently collected. This CNN is capable of predicting potential path directions with high confidence levels based on a single, uncalibrated spherical image. Experimental results demonstrate that the proposed framework outperforms competing ones in realistic applications.},
language = {en},
number = {6},
urldate = {2018-09-25},
journal = {Sensors},
author = {Ran, Lingyan and Zhang, Yanning and Zhang, Qilin and Yang, Tao},
month = jun,
year = {2017},
pages = {1341},
}
Downloads: 0
{"_id":"kL9DSqWg4u9ap4c8H","bibbaseid":"ran-zhang-zhang-yang-convolutionalneuralnetworkbasedrobotnavigationusinguncalibratedsphericalimages-2017","downloads":0,"creationDate":"2018-10-16T12:47:06.668Z","title":"Convolutional Neural Network-Based Robot Navigation Using Uncalibrated Spherical Images","author_short":["Ran, L.","Zhang, Y.","Zhang, Q.","Yang, T."],"year":2017,"bibtype":"article","biburl":"https://api.zotero.org/users/5038325/collections/87TDVNHG/items?key=OK1d4PeY6Vn5nhaS7wTquuuA&format=bibtex&limit=100","bibdata":{"bibtype":"article","type":"article","title":"Convolutional Neural Network-Based Robot Navigation Using Uncalibrated Spherical Images","volume":"17","issn":"1424-8220","url":"http://www.mdpi.com/1424-8220/17/6/1341","doi":"10.3390/s17061341","abstract":"Vision-based mobile robot navigation is a vibrant area of research with numerous algorithms having been developed, the vast majority of which either belong to the scene-oriented simultaneous localization and mapping (SLAM) or fall into the category of robot-oriented lane-detection/trajectory tracking. These methods suffer from high computational cost and require stringent labelling and calibration efforts. To address these challenges, this paper proposes a lightweight robot navigation framework based purely on uncalibrated spherical images. To simplify the orientation estimation, path prediction and improve computational efficiency, the navigation problem is decomposed into a series of classification tasks. To mitigate the adverse effects of insufficient negative samples in the “navigation via classification” task, we introduce the spherical camera for scene capturing, which enables 360◦ fisheye panorama as training samples and generation of sufficient positive and negative heading directions. The classification is implemented as an end-to-end Convolutional Neural Network (CNN), trained on our proposed Spherical-Navi image dataset, whose category labels can be efficiently collected. This CNN is capable of predicting potential path directions with high confidence levels based on a single, uncalibrated spherical image. Experimental results demonstrate that the proposed framework outperforms competing ones in realistic applications.","language":"en","number":"6","urldate":"2018-09-25","journal":"Sensors","author":[{"propositions":[],"lastnames":["Ran"],"firstnames":["Lingyan"],"suffixes":[]},{"propositions":[],"lastnames":["Zhang"],"firstnames":["Yanning"],"suffixes":[]},{"propositions":[],"lastnames":["Zhang"],"firstnames":["Qilin"],"suffixes":[]},{"propositions":[],"lastnames":["Yang"],"firstnames":["Tao"],"suffixes":[]}],"month":"June","year":"2017","pages":"1341","bibtex":"@article{ran_convolutional_2017,\n\ttitle = {Convolutional {Neural} {Network}-{Based} {Robot} {Navigation} {Using} {Uncalibrated} {Spherical} {Images}},\n\tvolume = {17},\n\tissn = {1424-8220},\n\turl = {http://www.mdpi.com/1424-8220/17/6/1341},\n\tdoi = {10.3390/s17061341},\n\tabstract = {Vision-based mobile robot navigation is a vibrant area of research with numerous algorithms having been developed, the vast majority of which either belong to the scene-oriented simultaneous localization and mapping (SLAM) or fall into the category of robot-oriented lane-detection/trajectory tracking. These methods suffer from high computational cost and require stringent labelling and calibration efforts. To address these challenges, this paper proposes a lightweight robot navigation framework based purely on uncalibrated spherical images. To simplify the orientation estimation, path prediction and improve computational efficiency, the navigation problem is decomposed into a series of classification tasks. To mitigate the adverse effects of insufficient negative samples in the “navigation via classification” task, we introduce the spherical camera for scene capturing, which enables 360◦ fisheye panorama as training samples and generation of sufficient positive and negative heading directions. The classification is implemented as an end-to-end Convolutional Neural Network (CNN), trained on our proposed Spherical-Navi image dataset, whose category labels can be efficiently collected. This CNN is capable of predicting potential path directions with high confidence levels based on a single, uncalibrated spherical image. Experimental results demonstrate that the proposed framework outperforms competing ones in realistic applications.},\n\tlanguage = {en},\n\tnumber = {6},\n\turldate = {2018-09-25},\n\tjournal = {Sensors},\n\tauthor = {Ran, Lingyan and Zhang, Yanning and Zhang, Qilin and Yang, Tao},\n\tmonth = jun,\n\tyear = {2017},\n\tpages = {1341},\n}\n\n","author_short":["Ran, L.","Zhang, Y.","Zhang, Q.","Yang, T."],"key":"ran_convolutional_2017","id":"ran_convolutional_2017","bibbaseid":"ran-zhang-zhang-yang-convolutionalneuralnetworkbasedrobotnavigationusinguncalibratedsphericalimages-2017","role":"author","urls":{"Paper":"http://www.mdpi.com/1424-8220/17/6/1341"},"downloads":0},"search_terms":["convolutional","neural","network","based","robot","navigation","using","uncalibrated","spherical","images","ran","zhang","zhang","yang"],"keywords":[],"authorIDs":[],"dataSources":["4hKjwEn3FuiBQWi8X"]}