請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/70718
完整後設資料紀錄
DC 欄位 | 值 | 語言 |
---|---|---|
dc.contributor.advisor | 洪一平 | |
dc.contributor.author | Pin-Hsin Lin | en |
dc.contributor.author | 林品忻 | zh_TW |
dc.date.accessioned | 2021-06-17T04:36:00Z | - |
dc.date.available | 2028-08-08 | |
dc.date.copyright | 2018-08-14 | |
dc.date.issued | 2018 | |
dc.date.submitted | 2018-08-08 | |
dc.identifier.citation | [1] Bârsan, I. A., et al., “Robust Dense Mapping for Large-Scale Dynamic Environments.” In: Proceedings of the IEEE International Conference on Robotics and Automation (ICRA). 2018.
[2] Girshick, R., et al., “Rich feature hierarchies for accurate object detection and semantic segmentation.” In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 2014. [3] Girshick, R., “Fast r-cnn.” In: Proceedings of the IEEE international conference on computer vision. 2015. [4] Ren, S., et al., “Faster r-cnn: Towards real-time object detection with region proposal networks.” In: Advances in Neural Information Processing Systems (NIPS). 2015. [5] Redmon, J., et al., “You only look once: Unified, real-time object detection.” In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. (CVPR) 2016. [6] Liu, W., et al., “Ssd: Single shot multibox detector.” In: European Conference on Computer Vision (ECCV). Springer. Cham. 2016. [7] Dai, J., et al., “Instance-aware semantic segmentation via multi-task network cascades.” In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 2016. [8] Pinheiro, P. O., et al., “Learning to refine object segments.” In: European Conference on Computer Vision (ECCV). Springer. Cham. 2016. [9] Intel RealSense Depth Camera SR300. https://click.intel.com/intelrealsense-developer-kit-featuring-sr300.html [10] Weeview Stereo Camera. https://www.esentra.com.tw/product/sid-3d-camera/ [11] Kähler, O., et al., “Very high frame rate volumetric integration of depth images on mobile devices.” IEEE Transactions on Visualization and Computer Graphics (TVCG). 2015. [12] Pinheiro, P. O., et al., “Learning to segment object candidates.” In: Advances in Neural Information Processing Systems (NIPS). 2015. [13] Geiger, A., et al., “Stereoscan: Dense 3d reconstruction in real-time.” In: IEEE Intelligent Vehicles Symposium (IV). 2011. [14] Geiger, A., et al., “Efficient large-scale stereo matching.” Asian Conference on Computer Vision (ACCV). Springer. 2010. [15] Lin, T. Y., et al., “Microsoft coco: Common objects in context.” In: European Conference on Computer Vision (ECCV). Springer. Cham. 2014. [16] Wu, C., “Towards linear-time incremental structure from motion.” In: International Conference on 3D Vision (3DV). 2013. [17] Mur-Artal, R., et al., “ORB-SLAM: a versatile and accurate monocular SLAM system.” IEEE Transactions on Robotics. 2015. [18] Newcomben, R. A., et al., “KinectFusion: Real-time dense surface mapping and tracking.” In: IEEE International Symposium on Mixed and augmented reality (ISMAR). 2011. [19] Curless, B., et al., “A volumetric method for building complex models from range images.” In: Proceedings of conference on Computer graphics and interactive techniques. ACM. 1996. [20] GoPro, GoPro Hero 4, https://zh.shop.gopro.com/International/cameras. [21] Hu, Y. T., et al., “Maskrnn: Instance level video object segmentation.” In: Advances in Neural Information Processing Systems (NIPS). 2017. | |
dc.identifier.uri | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/70718 | - |
dc.description.abstract | 即時定位與地圖構建(SLAM)是一種用來解決機器人自我定位問題的方法。近年來這個問題越來越多人在討論及研究。一般而言,SLAM的方法都假設是在靜態環境下做定位。但是在現實生活中是個動態的環境,例如有行人或是車輛在走動等等。如果我們使用在動態物體身上的特徵點來做定位,會影響到定位的準確度。因此我們提出了透過結合兩種深度學習方法的優點,來去除屬於動態物體的特徵點,用剩餘的畫面來做定位的方法。在本篇論文中,我們著重在探討目前新興的深度學習物體切割方法,並且研究移除動態物體特徵點對定位結果的影響。最後我們的方法能夠增加每一幀中偵測到可能會動的物體的召回率,並且在定位同時判斷這些物體的運動狀態後,將動態物體穩定地去除以提升定位準確度。 | zh_TW |
dc.description.abstract | Simultaneous Localization and Mapping (SLAM) is a solution of robotic ego-positioning problem, which is more and more popular nowadays. Normally, it was assumed that the SLAM technique can only be performed in static environments. However, we are often in a dynamic environment, such as those containing other vehicles or pedestrians. Using features on dynamic objects to do SLAM will influence the positioning accuracy. Therefore, we proposed a method to use images after dynamic object segmentation during SLAM by combining the advantages of two deep-learning-based segmentation methods. In this paper, we focus on investigating the state-of-the-art deep-learning-based segmentation methods and the impact of dynamic object segmentation on SLAM. Our method can first increase the recall of detecting potential moving objects in each frame and neglect dynamic objects robustly to improve the positioning accuracy. | en |
dc.description.provenance | Made available in DSpace on 2021-06-17T04:36:00Z (GMT). No. of bitstreams: 1 ntu-107-R05944008-1.pdf: 3273356 bytes, checksum: fe555024d31e43c61960e42ffc900f00 (MD5) Previous issue date: 2018 | en |
dc.description.tableofcontents | 口試委員會審定書 I
誌謝 II 中文摘要 III ABSTRACT IV CONTENTS V LIST OF FIGURES VII LIST OF TABLES IX Chapter 1 Introduction 1 1.1 Motivation 1 1.2 Deep-learning-based Object Segmentation 2 1.3 Type of Cameras in SLAM 4 Chapter 2 Related Work 6 2.1 SLAM System 6 2.2 Deep-learning-based Object Segmentation 8 Chapter 3 Dynamic SLAM Based on Deep Learning 10 3.1 System Overview 10 3.1.1 Deep-learning-based Segmentation Method Combination 10 3.1.2 Dynamic Object Segmentation 12 3.2 System Details 13 3.2.1 Deep-learning-based Segmentation Method Combination 13 3.2.2 Sparse Scene Flow and Masked Scene Flow 13 3.2.3 Robust Visual Odometry and Object Motion Detection 14 3.2.4 Depth Computation and Static Map Reconstruction 16 Chapter 4 Experiments 18 4.1 Deep-learning-based Segmentation 18 4.1.1 Experiment Purpose 18 4.1.2 Experiment Evaluation 19 4.1.3 Experiment Result 20 4.2 SLAM in Dynamic Environments 21 4.2.1 Experiment Equipment 21 4.2.2 Experiment Evaluation 22 4.2.3 Experiment Scenario 1: Environments with Many Small and Medium Dynamic Objects 24 4.2.4 Experiment Scenario 2: Environments with One Large Dynamic Object 29 4.3 Summary 34 Chapter 5 Conclusion 36 Chapter 6 Future Work 37 REFERENCE 39 | |
dc.language.iso | en | |
dc.title | 在動態環境中使用深度學習之基於物體切割的即時定位與地圖構建 | zh_TW |
dc.title | SLAM with Object Segmentation in Dynamic Environments Using Deep Learning | en |
dc.type | Thesis | |
dc.date.schoolyear | 106-2 | |
dc.description.degree | 碩士 | |
dc.contributor.oralexamcommittee | 莊榮宏,陳冠文,邱志義,石勝文 | |
dc.subject.keyword | 即時定位與地圖構建,深度學習,物體分割,動態環境,戶外環境, | zh_TW |
dc.subject.keyword | simultaneous localization and mapping,deep learning,object segmentation,dynamic environment,outdoor environment, | en |
dc.relation.page | 40 | |
dc.identifier.doi | 10.6342/NTU201802702 | |
dc.rights.note | 有償授權 | |
dc.date.accepted | 2018-08-09 | |
dc.contributor.author-college | 電機資訊學院 | zh_TW |
dc.contributor.author-dept | 資訊網路與多媒體研究所 | zh_TW |
顯示於系所單位: | 資訊網路與多媒體研究所 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-107-1.pdf 目前未授權公開取用 | 3.2 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。