Font Size: a A A

Research And Application Of 3D Visualization Of Urban Real Estate Based On Cesium Framework

Posted on:2021-05-01Degree:MasterType:Thesis
Country:ChinaCandidate:C X WangFull Text:PDF
GTID:2370330614970123Subject:Software engineering
Abstract/Summary:PDF Full Text Request
With the rapid development of cities and the advent of the Internet era,smart cities have become a hot research topic in urban construction.Real estate is one of the basic elements of the city,and the three-dimensional visualization of urban real estate is one of the important basic functions of smart city applications.In the process of building a smart city,rapid three-dimensional indoor modeling of micro real estate and large-scale three-dimensional visualization of Web scenes have always been the technical problems that limit the application of smart cities in 3D scenes.In response to the above two requirements,combined with Web three-dimensional visualization and deep learning technologies,this paper has conducted theoretical and technical research on the three-dimensional rapid modeling of microscopic real estate indoor scenes and large-scale macroscopic urban real estate visualization.The main work and results of this article are as follows:Firstly,an indoor scene modeling method based on floor plan instance segmentation is proposed.Aiming at the problems of high cost and low efficiency of 3D modeling of indoor scenes,through the research and application of deep learning,the Mask R-CNN algorithm is applied to the instance segmentation of the floor plan,and combined with image registration to quickly obtain spatial layout information from two-dimensional images.Finally,the three-dimensional modeling of the indoor scene is realized through Three JS and the furniture model library.Secondly,the Web 3D visualization technology of the large-scale 3D model of real estate is studied.Based on Cesium,Image terrain data and oblique photography model data,which are two major large-scale data,is loaded and 3D rendered,and the fineness of the model is dynamically adjusted according to FPS to improve the user experience.Thirdly,design and implement a 3D visualization system of urban real estate based on Web.Based on the indoor scene 3D modeling technology and large-scale 3D scene rendering method proposed in this article,a 3D visualization system of urban real estate is realized.Users can roam the city through this system,can view the indoor scene of the real estate,and can also calculate the space Dynamic loading of external data,etc.Experiments show that the technical solution proposed in this paper can better solve the problems of rapid 3D modeling visualization of microscopic indoor scenes and visualization of large-scale macroscopic 3D real estate data,and has certain significance for promoting the application of urban 3D visualization technology in smart cities.
Keywords/Search Tags:smart city, Cesium, visualization, indoor scene, deep learning
PDF Full Text Request
Related items