logo资料库

Radu Bogdan Rusu's Phd Thesis（PCL在这篇论文中诞生）.pdf

发布时间：2022-05-29 发布人：admin 分类：说明书 资料大小：59.98M 资料格式：pdf 举报 版权申诉
第1页 / 共284页
第2页 / 共284页
第3页 / 共284页
第4页 / 共284页
第5页 / 共284页
第6页 / 共284页
第7页 / 共284页
第8页 / 共284页
资料共284页，剩余部分请下载后查看
    Abstract
    Contents
    List of Figures
    List of Tables
    List of Algorithms
    List of Symbols and Notations
    Introduction
    Why ``3D'' Semantic Perception?
    Computational Problems
    Publications
    Thesis Outline and Contributions
    Semantic 3D Object Mapping Kernel
    3D Map Representations
    Data Acquisition
    Data Representation
    Summary
    Mapping System Architectures
    3D Point Feature Representations
    The ``Neighborhood'' Concept
    Filtering Outliers
    Surface Normals and Curvature Estimates
    Point Feature Histograms (PFH)
    Fast Point Feature Histograms (FPFH)
    Feature Persistence
    Related Work
    Summary
    From Partial to Complete Models
    Point Cloud Registration
    Data Resampling
    Related Work
    Summary
    Clustering and Segmentation
    Fitting Simplified Geometric Models
    Basic Clustering Techniques
    Finding Edges in 3D Data
    Segmentation via Region Growing
    Application Specific Model Fitting
    Summary
    Mapping of Indoor Environments
    Static Scene Interpretation
    Heuristic Rule-based Functional Reasoning
    Learning the Scene Structure
    Exporting and Using the Models
    Summary
    Surface and Object Class Learning
    Learning Local Surface Classes
    Generating Training Data
    Most Discriminative Feature Selection
    Supervised Class Learning using Support Vector Machines
    Fast Geometric Point Labeling
    Global Fast Point Feature Histograms for Object Classification
    Summary
    Parametric Shape Model Fitting
    Object Segmentation
    Hybrid Shape-Surface Object Models
    Summary
    Applications
    Table Cleaning in Dynamic Environments
    Real-time Collision Maps for Motion Re-Planning
    Semantic Interpretation of 3D Point Cloud Maps
    System Evaluation
    Summary
    Identifying and Opening Doors
    Detecting Doors
    Detecting Handles
    System Evaluation
    Summary
    Real-time Semantic Maps from Stereo
    Leaving Flatland Mapping Architecture
    Visual Odometer
    Spatial Decomposition
    Polygonal Modeling
    Merging and Refinement
    Semantic Labeling
    3D Mapping Performance
    Semantic Map Usage and Applications
    Hybrid Model Visualizations
    Motion Planning for Navigation
    Summary
    Conclusion
    3D Geometry Primer
    Euclidean Geometry and Coordinate Systems
    Distance Metrics
    Geometric Shapes
    Sample Consensus
    Machine Learning
    Support Vector Machines
    Conditional Random Fields
    Bibliography
    Institut für Informatik der Technischen Universität München Semantic 3D Object Maps for Everyday Manipulation in Human Living Environments Dissertation Radu Bogdan Rusu
    TECHNISCHE UNIVERSITÄT MÜNCHEN Institut für Informatik Semantic 3D Object Maps for Everyday Manipulation in Human Living Environments Radu Bogdan Rusu Vollständiger Abdruck der von der Fakultät für Informatik der Technischen Universität München zur Erlangung des akademischen Grades eines Doktors der Naturwissenschaften (Dr. rer. nat.) genehmigten Dissertation. Vorsitzender: Univ.-Prof. Nassir Navab, Ph.D. Prüfer der Dissertation: 1. Univ.-Prof. Michael Beetz, Ph.D. 2. Prof. Kurt Konolige, Ph.D., Stanford University, Palo Alto, USA 3. Prof. Gary Bradski, Ph.D., Stanford University, Palo Alto, USA Die Dissertation wurde am 14.07.2009 bei der Technischen Universität München eingereicht und durch die Fakultät für Informatik am 18.09.2009 angenommen.
    Abstract ENVIRONMENT models serve as important resources for an autonomous robot by provid- ing it with the necessary task-relevant information about its habitat. Their use enables robots to perform their tasks more reliably, flexibly, and efficiently. As autonomous robotic platforms get more sophisticated manipulation capabilities, they also need more expressive and comprehensive environment models: for manipulation purposes their models have to in- clude the objects present in the world, together with their position, form, and other aspects, as well as an interpretation of these objects with respect to the robot tasks. This thesis proposes Semantic 3D Object Models as a novel representation of the robot’s operating environment that satisfies these requirements and shows how these models can be automatically acquired from dense 3D range data. The thesis contributes in two important ways to the research area acquisition of environment models. The first contribution is a novel framework for Semantic 3D Object Model acquisition from Point Cloud Data. The functionality of this framework includes robust alignment and integra- tion mechanisms for partial data views, fast segmentation into regions based on local surface characteristics, and reliable object detection, categorization, and reconstruction. The computed models are semantic in that they infer structures in the data that are meaningful with respect to the robot task. Examples of such objects are doors and handles, supporting planes, cupboards, walls, or movable smaller objects. The second key contribution is point cloud representations based on 3D point feature his- tograms (3D-PFHs), which model the local surface geometry for each point. 3D-PFHs dis- tinguish themselves from alternative 3D feature representations in that they are very fast to compute, robust against variations in pose and sampling density, and cope well with noisy sensor data. Their use substantially improves the quality of the Semantic 3D Object Models acquired, as well as the speed with which they are computed. 3D-PFHs come with specific software tools that allow for the learning of surface characteristics based on their underlying geometry, the assembly of most distinctive 3D points from a given cloud, as well as limited view-invariant correspondence search for 3D registration. III
    The contributions presented in this thesis have been fully implemented and empirically evaluated on different robots performing different tasks in different environments. The first demonstration relates to the problem of cleaning tables by disposing the objects on them into a garbage bin with a personal robotic assistant in the presence of humans in its working space. The framework for Semantic 3D Object Model acquisition is demonstrated and used to con- struct dynamic 3D collision maps, annotate the surrounding world with semantic labels, and extract object clusters supported by tables in real-time performance. The second demonstra- tion presents an on-the-fly model acquisition system for door and handle identification from noisy 3D point cloud maps. Experimental results show good robustness in the presence of large variations in the data, without suffering from the classical under or over-fitting problems usually associated with similar initiatives based on machine learning classifiers. The third ap- plication example tackles the problem of real-time semantic mapping of indoor environments with different kinds of terrain classes, such as walkways and stairs, for the navigation of a six-legged robot with terrain-specific walking modes.
    Contents Abstract Contents List of Figures List of Tables List of Algorithms List of Symbols and Notations 1 Introduction 1.1 Why “3D” Semantic Perception? . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Computational Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3 Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4 Thesis Outline and Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I Semantic 3D Object Mapping Kernel 2 3D Map Representations . . . 2.1 Data Acquisition . . 2.2 Data Representation . . 2.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Mapping System Architectures 4 3D Point Feature Representations V III V IX XVII XIX XXI 1 3 4 6 11 15 17 17 23 28 31 37
    . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1 The “Neighborhood” Concept 4.2 Filtering Outliers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3 Surface Normals and Curvature Estimates . . . . . . . . . . . . . . . . . . . 4.4 Point Feature Histograms (PFH) . . . . . . . . . . . . . . . . . . . . . . . . 4.5 Fast Point Feature Histograms (FPFH) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.6 Feature Persistence . . 4.7 Related Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.8 Summary . . . . . 5 From Partial to Complete Models 5.1 Point Cloud Registration . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2 Data Resampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3 Related Work . . 5.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Clustering and Segmentation 6.1 Fitting Simplified Geometric Models . . . . . . . . . . . . . . . . . . . . . . 6.2 Basic Clustering Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3 Finding Edges in 3D Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.4 Segmentation via Region Growing . . . . . . . . . . . . . . . . . . . . . . . 6.5 Application Specific Model Fitting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.6 Summary . . . . . . . . . II Mapping of Indoor Environments 39 42 45 50 57 61 65 67 69 69 77 81 83 85 86 88 90 91 93 96 97 7 Static Scene Interpretation 99 7.1 Heuristic Rule-based Functional Reasoning . . . . . . . . . . . . . . . . . . 101 7.2 Learning the Scene Structure . . . . . . . . . . . . . . . . . . . . . . . . . . 111 7.3 Exporting and Using the Models . . . . . . . . . . . . . . . . . . . . . . . . 117 . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 7.4 Summary . . . . . . . . . 8 Surface and Object Class Learning 123 8.1 Learning Local Surface Classes . . . . . . . . . . . . . . . . . . . . . . . . . 125 8.1.1 Generating Training Data . . . . . . . . . . . . . . . . . . . . . . . . 127 8.1.2 Most Discriminative Feature Selection . . . . . . . . . . . . . . . . . 130 Supervised Class Learning using Support Vector Machines . . . . . . 133 8.1.3
    分享到：
    赞 收藏

    相关推荐

    资料库

    开发技术

    共收录19306份资料，累计14个分类，关注成员有19位，主要包括：Actionscript,Javascript,PHP,VB,Perl,Delphi,Python,C#,C,Java,C++,Web开发,硬件开发,其它

    热门标签

    Actionscript Javascript PHP VB Perl Delphi Python C# C Java C++ Web开发 硬件开发 其它

    最新资料