The Analysis Of 3D Scene Data Parallel Processing Strategy And Model Simplification

After research of Big-data,Internet,and Cloud-computing on GIS is embedded and extensive,it allows us to enter the era of smart city by the digital city era.Meanwhile,the functional requirements of data analysis are not only satisfied with the visual display for certain information,but the higher requirements for visualization of knowledge,certainty of decision,and real-time of acquisition.The 3D GIS technology is facing new development opportunities and challenges due to these requirements.With the rapid increase of data in TB(terabyte)or GB(gigaByte)scale and the real-time upgrading of complexity,it is difficult to meet the requirements of collecting,storing,retrieving,analyzing,mining,and visualizing in an acceptable time for existing database management software and traditional data processing technology.It is difficult to guarantee the stable and smooth experience of user interaction due to the large amount of calculation and time-consuming in the drawing,query,and display for3 D data.In this paper,based on the above background,we thoroughly analyzed and deeply extended the organization method for 3D city model data and the LOD model,which deepens the spatial analysis function of 3D model.The main works and research results are as follows:(1)it aims at dealing with the complex and large data by parallel processing method using heterogeneous architecture.The approach takes GPU memory storage capacity into account,also discusses the basic principle of the heterogeneous structure on parallel operational framework,and heterogeneous schema data in the load measuring and data transmission are discussed and analyzed.(2)A dynamic continuous LOD simplification method for integrated terrain and terrain model is established,which can control the quality and accuracy of output model and upgrade the classical model reduction method.(3)After using semantic recognition,we adopts the theory of spatial cognition and city morphology.Firstly,we build constrained Delauney triangulation with surface boundary contour point's automatic extraction,and we constrain the triangulation by the road network.Then we apply elevation,orientation and area threshold and topology for classifying the global scene based on Gestalt theory to fit with the characteristics of urban morphology.