Multivariate Data Based Visualization Of "unicorn Island" 3d Scene

With the increasing demand for "Digital City",virtual reality and geographic information technology are also making continuous breakthroughs,and the research on 3D visualization of cities is booming all over the world.Therefore,new methods and technologies for constructing 3D visualization also emerge at the right moment.In the current mainstream city 3D modeling technology,the traditional artificial modeling technology,although the model precision is high,but the efficiency is low,the cost of labor force;The cost of automatic modeling based on laser scanner is too high.When using photogrammetry to build the model,the Unmanned Aerial Vehicle is often restricted by air traffic control and other conditions,while the camera cannot shoot the complete feature,which makes the model accuracy and integrity cannot be guaranteed.Therefore,it is difficult for a single modeling technology to accurately and quickly restore the real 3D scene.This study in Cheng Du Tian Fu new area unicorn island as an example,according to the characteristics of the different features in the study area,the structure and the precision requirement,by combining multiple views photogrammetry with production softwares 3 DMAX technology advantages,and the optimization of point cloud data processing,summarizes a set of high accuracy,high efficiency and saving the cost of city three dimensional space construction method,and the 3D visualization in the practical application to do the research in the field of city management and so on,the main achievements are as follows:(1)A method based on multi-view photogrammetry and 3DMAX modeling was proposed.For most and complex features in the city,the model is constructed quickly by using multi-view close-up photogrammetry.Due to occlusion or other environmental factors,it is impossible to obtain a complete ground object image,which results in partial loss of ground object model or structural rules.3DMAX is used to repair and construct ground object based on remote sensing image,which makes the model more complete and effectively saves the modeling time.(2)Model optimization for point cloud data generates high-precision point cloud model.Based on multiple views of photogrammetry for point cloud are scattered,sparse and city 3D scene visualization of model data accuracy is higher,in order toimprove the model accuracy,reduce the financial burden on computer operation,using PMVS(patch-based multi-view stereo)point cloud dense algorithm and grid hole repair algorithm,the sparse to dense point cloud and empty repair process,the resulting high precision point cloud model.(3)Realized 3D scene visualization and spatial analysis.With the help of visualization technology,the urban elements constructed by multiple data are expressed in the same scene to complete the construction of 3d scene,and the spatial technology of GIS is integrated into the 3d urban scene to realize the spatial analysis and planning analysis of 3D visualization.This paper discusses the concrete application of three-dimensional visualization of urban space,and shows the practical application value of 3D urban scene in real life.