Research On Extraction Method Of Accurate 3D Buffer Surface Based On Map Algebraic Distance Field

In recent years,with the continuous advancement of spatial data acquisition methods,spatial data presents multi-dimensional characteristics,and three-dimensional GIS has become one of the research hotspots in the field of GIS.Most of the existing 3D GIS research focuses on 3D spatial data acquisition,3D spatial data model,3D spatial data visualization,etc.The research related to 3D spatial analysis needs to be further developed.Spatial analysis is the biggest feature of the GIS system that is different from other systems.Buffer analysis is one of the most basic and important functions in spatial analysis.The results are often used as basic data for other spatial analysis.At present,the buffer analysis algorithm in 2D GIS is quite mature,but the buffer analysis algorithm in 3D GIS needs to be improved in both accuracy and efficiency compared to 2D.Therefore,the study of efficient and accurate 3D buffer analysis algorithm is It is very necessary.The results of the vector buffer analysis algorithm are more accurate,which also means that the space algorithm has a higher complexity.There are a lot of spatial geometric intersection operations in the process,which requires a lot of special processing.The raster algorithm is based on distance transformation.The idea and process are relatively simple,but the accuracy of the algorithm and the visualization of the results are affected by the grid resolution,and they have high requirements on computer hardware.Today,the computer's storage level has been greatly improved,GPU parallel computing technology is quite mature,and raster algorithms are no longer limited by the level of computer hardware.Under this condition,this paper proposes an accurate 3D buffer surface extraction method based on map algebraic distance field supported by vector grid and data model.This research first extended the two-dimensional map algebraic distance transformation algorithm to three-dimensional,and used the actual distance instead of the template to transfer the distance,thereby further improving the accuracy of the three-dimensional distance field.Then,the concept of boundary voxels and boundary voxels equivalent is proposed in this study,and the topological relationship of boundary voxel pairs is defined.The surface feature points of the three-dimensional buffer surface and the topological relationship between them are extracted by the boundary voxel pairs and their topological relationships.Finally,the idea of moving cube algorithm is introduced to complete the reconstruction of the three-dimensional buffer surface from a local perspective.This study verified the accuracy of the algorithm from three aspects: buffer shape,buffer distance,and voxel resolution.The results show that the method has high accuracy and controllability.The higher the voxel resolution,the higher the accuracy and buffer distance has little effect on algorithm accuracy.The efficiency of the proposed algorithm is analyzed from the time complexity.The results show that the efficiency of this method is only related to the number of voxels after voxelization in3 D space,and has nothing to do with the shape,number and complexity of the target space entities.Finally,the paper discusses the application prospects of the algorithm in 3D terrain analysis and medical field.