Study On Key Techniques Of Automatically Generating 3D Indoor Map Data Based On CAD Floor Plan

With the rise of "indoor location services",indoor maps,as an important basis for location services,have attracted wide attention.The three-dimensional indoor map has become the preferred expression of indoor maps because of its real sense of enlightenment and better user experience.Nowadays,three-dimensional indoor maps are mainly from user-uploaded building plans and post-processing or collecting data by use of some kind of hardware and then post-processing.Because the automation level of these two cartographic modes is not high,it makes the contraction between cartography efficiency and demand more prominent when large quantities of cartography are needed.How to improve the cartography efficiency of 3D indoor maps and create a semi-automatic or automated technology system has become one of the key issues to be solved urgently in the field of indoor location services.This article focuses on a fundamental issue in the field of three-dimensional indoor maps: automatic data acquisition and generation.Considering that the CAD software in the domestic construction field has been promoted for many years and a large number of CAD drawings have been accumulated,this paper proposes a method for automatically generating 3D indoor map data based on a CAD floor plan.The following is the main work done in this article:1.In view of the lack of robustness,efficiency and insufficient extraction parameters for the existing CAD drawing analysis methods,firstly,we designed an algorithm that merges adjacent wall and gland parallel wall lines to preprocess the drawings,which can improve the fault tolerance of the search loop algorithm;then a new drawing parsing strategy was proposed around the creation and use of custom data structures-semantic polygons.Based on this,several algorithms were designed to improve the extraction efficiency,accuracy and richness of conventional building components-doors and windows.And extraction detailed parameters of complex building components like staircase and elevator are also implemented that make them support parametric modeling during the three-dimensional reconstruction phase;finally,we use the improved search loop algorithm to extract functional zone boundaries and identify functional area type by expanding the traditional methods based on semantic annotation to obtains the spatial pattern information of the entire floor.2.In view of the lack that some edges of the vector path which generated by existing algorithm are so near to wall lines,this paper proposed a new algorithm to overcome this shortcoming.vector indoor road network generation algorithm is proposed.Because the road network generated by this algorithm is clear and simple,the storage space needed is also lower.In addition,in order to make the results more versatile,this paper also designed a method to convert the indoor road network results into a universal road network storage format—the undirected weighted graph.3.To rebuild the indoor scene and integrated storage of 2/3D indoor map data,this paper presents a parameterized modeling algorithm for building components and a scene assembly method to rebuild indoor scene based on the interface named ArcEngine.And then we designed a 2/3D data integrated storage model for Geodatabase,which come from the mainstream GIS platform named ArcGIS.4.Establishing a system to implement the strategies and methods for generating 3D indoor map data based on CAD floor plan.We established a three-dimensional indoor map data generation system using.NET language and ArcEngine component technology.The system uses CAD files in.DXF format as the data source and with the helper of the guide,it’s easy for you to generate a two-dimensional indoor map(.mxd)and a multi-floor three dimensional indoor map data(.gdb)from one standard layer drawing.