Southwest Jiaotong University School of Mathematics


ag网站亚游登录  >  学术科研  >  信息与计算科学系  >  正文

学术报告 Stochastic and Combinatorial Reasoning for 3D Indoor Models

ag网站亚游登录:信息与计算科学系   作者:信息与计算科学系     日期:2019-04-01 17:46:24   点击数:  


 Stochastic and Combinatorial Reasoning for 3D Indoor Models Based on Sparse Data

报告人 :Prof. Dr. rer.nat. Lutz Plümer







     Graphical models are probabilistic networks which have widely been used as means of deriving probabilistic hypotheses from noisy data. We adapt and general a special case, namely conditional linear Gaussian (CLG) networks, in order to derive 3D indoor Models from sparse data. We use as input only the floor plan of the building, and the area, the numbers and the type of use of the individual rooms and beyond that the position of the windows. Such data are widely available, and the presentation will show how it is possible to derive reasonable hypotheses with high accuracies and probabilities from such data. The point is that we exploit the regular structure of buildings and a priori knowledge, that is distributions of the essential model parameters. We use methods from the field of combinatorics, namely constraint propagation, and adjustment theory, namely Gau? Markov models, Kalman filters, mixed Gaussian distributions.


Prof. Dr. rer.nat. Lutz Plümer

International Dean of the Faculty of Geosciences and Environmental Engineering

South West Jiaotang University

Former Managing Director of the Institute for Cartography and Geoinformation and
Chairman of the GIS Centre of the University of Bonn

Senator of the University of Bonn,
Chairman of the Finding Commission for the University Council
Member of the Bavarian Academy of Science (German Geodetic Commission)

  • 1968 – 1975 University of Hamburg (Germany) / Educational Science, Philosophy, Sociology

  • 1982 – 1986 University of Dortmund (Germany) / Computer Science / Diploma
    (Grant of the German National Scholarship Foundation)

  • 1986 – 1990 Research Associate at the Institute of Computer Science VI, University of Dortmund (Germany)

  • 1989 Promotion at the University of Dortmund (Germany), summa cum laude

    Thesis: Termination Proofs for Logic Program

  • 1991 – 1997 Research Associate at the Institute of Computer Science III, University of Bonn (Germany)

  • 1994 Habilitation at the University of Bonn

    Thesis: Automatic Verification of Parallel Logic Programs

  • 1997 – 1999 Professor at the Institute for Environmental Sciences / University of Vechta (Germany)

  • 1999 - 2016 Full Professor for Geoinformation, Institute of Geodesy and Geoinformation / University of Bonn (Germany)

  • 2018 -  International Dean of the Faculty of Geosciences and Environmental Engineering of  South West Jiaotang University

Research fields:

  • Geoinformation

  • Expert Systems

  • Geometric-Topological Integrity

  • Standardization of City Models -CityGml

  • Semantic Reconstruction of 3D Buildings from 3D Point Clouds

  • Machine Learning for Geoinformation

  • Applications of Machine Learning in Agriculture



XML 地图 | Sitemap 地图