首页--天文学、地球科学论文--测绘学论文--大地测量学论文--电磁波测距和基线测量论文--光电测距论文

城郊场景三维点云建筑识别

摘要第6-7页
Abstract第7页
Acknowledgements第9-12页
Chapter 1 Introduction第12-22页
    1.1 Motivation第12-13页
    1.2 Literature review第13-21页
        1.2.1 Point cloud feature extraction第13-18页
        1.2.2 Point cloud classification第18-21页
    1.3 Problem description第21页
    1.4 Contributions and Structures第21-22页
Chapter 2 Methodology第22-46页
    2.1 Workflow of building extraction from point clouds of suburban scenes第22-24页
    2.2 Point cloud downsampling and outlier removal第24-25页
        2.2.1 Point cloud downsampling using approximate voxel grid filter第24页
        2.2.2 Ground point removal using Digital Terrain Model (DTM) filter第24页
        2.2.3 Outlier removal using statistical outlier filter第24-25页
    2.3 Local feature extraction based on geometry第25-36页
        2.3.1 Local 3D shape features descriptor第25-26页
        2.3.2 Fast Point Feature Histograms (FPFH) descriptor第26-31页
        2.3.3 Signature of Histograms of Orientations (SHOT) descriptor第31-36页
    2.4 Supervised classification based on extracted features第36-43页
        2.4.1 K-nearest neighbor (k-NN) classifier第36-39页
        2.4.2 Discriminant Analysis (DA) classifier第39-41页
        2.4.3 Random forest(RF)classifier第41-43页
    2.5 Evaluation of predicted label第43-46页
        2.5.1 Frequently used evaluation criterion第43-45页
        2.5.2 Cross validation第45-46页
        2.5.3 Out of bag error第46页
Chapter 3 Experiments and results第46-107页
    3.1 Experiment data第46-56页
        3.1.1 Terrestrial laser scanning (TLS) point clouds第46-48页
        3.1.2 Arial photogrammetric point clouds第48-49页
        3.1.3 Ground truth labelling第49-56页
    3.2 Results of point cloud outlier removal and downsampling第56-63页
        3.2.1 Results of point cloud downsampling第56-59页
        3.2.2 Results of ground point removal第59-61页
        3.2.3 Results of outlier removal第61-63页
    3.3 Results of local feature extraction第63-76页
        3.3.1 Results of local 3D shape features descriptor第63-66页
        3.3.2 Results of norm estimation第66-70页
        3.3.3 Results of Fast Point Feature Histograms (FPFH) descriptor第70-73页
        3.3.4 Results of Signature of Histograms of Orientations (SHOT) descriptor第73-76页
    3.4 Results of Classification第76-107页
        3.4.1 Influence of classifier parameter selection第76-84页
        3.4.2 Influence of size of normal estimation neighborhood第84-88页
        3.4.3 Influence of size of feature extraction search area第88-94页
        3.4.4 Influence of extracting different features第94-96页
        3.4.5 Influence of classifier selection第96-97页
        3.4.6 Influence of combining with RGB value第97-98页
        3.4.7 Best results of classification第98-107页
Chapter 4 Conclusion, Discussion and outlook第107-109页
    4.1 Conclusion and discussion第107-108页
    4.2 Outlook第108-109页
List of Figures第109-113页
List of Tables第113-114页
Bibliography第114-116页

论文共116页,点击 下载论文
上一篇:基于标志点的近景摄影测量系统若干关键技术研究
下一篇:土地覆盖变化信息局域化精度估计方法研究