Research On 3d Building Modeling Method Based On GF-7 Satellite Data

For a long time,urban planning has been in a two-dimensional form,but the two-dimensional management model can no longer meet the growing needs of urban planning,so concepts such as smart cities,real-time 3D,and city information modeling have emerged,which effectively promote the informatization,three-dimensionality,and intelligence of cities.As the main element of real-time 3D and city information modeling,how to quickly and accurately obtain the 3D model of buildings is currently a research hotspot.To achieve the rapid construction of 3D models of buildings,this study selected high-resolution satellite images as the data source,combined with vector data from house surveys and 3D point cloud data from Li DAR,and tried to design and implement an experimental plan from high-resolution satellite data to the 3D model of urban buildings,and evaluated the accuracy of the results.The main work and conclusions of this paper are as follows:(1)In view of the low accuracy of the existing DEM data in urban DSM construction,this study uses a two-step method to generate urban DSM.Firstly,based on the semi-global algorithm and using high-resolution satellite data,Tianditu images,and open-source DEM,DSM was initially constructed.Then,an improved filtering algorithm based on region growing was used to remove ground objects from the DSM and generate a high-precision DEM.Finally,this DEM was used to replace the open-source DEM,achieving the reconstruction of DSM.Compared with the initially constructed DSM,the DSM generated by the high-precision DEM reduced the average error from 0.44 meters to 0.12 meters.(2)In view of the problem of large height errors in building extraction based on high-resolution satellite data,this study uses DSM and DEM to construct n DSM,fits the vector data from house surveys,and uses a zoning statistical method to extract feature parameters that can characterize building height.The optimal value of building height was then selected through error calculation.The results showed that the mean errors in the maximum and average values of building height parameters were the smallest,and the optimal value was used to extract building heights in the experimental area.(3)To serve the construction of city information models,this study uses the parameterized 3D modeling idea,combined with terrain,building 3D modeling,and texture mapping methods,to build a complete process from data reading,3D modeling,texture mapping,model output,and display,and achieve the 3D visualization of the terrain and buildings in the research area.This lays a technical foundation for the construction of the framework model and real-time model in the city information model.