Research And Application Of Large-scale 3D Scalar Field Data Visualization

3D scalar data visualization is an important research direction of computer graphics.Compared with the traditional two-dimensional visualization methods,three-dimensional visualization can not only fully display the distribution of data in three-dimensional space and reflect the full picture of the data field,but also achieve more realistic rendering effect.Thus,it can better meet the needs of different application scenarios.With the rapid development of modern science and technology,the means of data collection are becoming more and more diversified,which leads to a sharp increase in the amount of data collected by observation.The scientific visualization of large amounts of data needs sufficient memory and computing power to ensure the real-time performance of 3D rendering and rendering,which is a very challenging problem.Taking geological data as the object,this paper makes an in-depth study on the real-time visualization of 3D scalar data,which involves data compression,data switching,data modeling and so on.The main results are as follows:(1)Geological scalar data have the characteristics of multi-scale,multi-direction and local variation.In this paper,a volume data compression algorithm based on discrete sampling is proposed,which is graded according to the discrete degree of numerical distribution of geological data in different sampling intervals,so as to realize data compression.Ray casting algorithm is used to draw the compressed volume data.The experimental results show that the algorithm has high data compression ratio and low loss of graphics accuracy,and can meet the requirements of real-time rendering while maintaining reliability.(2)The direct internal and external scheduling of large-scale geological scalar data blocks will lead to the delay of visualization results.In this paper,a real-time switching algorithm of logging attributes based on subspace learning is proposed.firstly,logging attributes are classified based on correlation analysis,and then the basis vectors and their coefficients are obtained through intra-class subspace learning,based on which the mapping model between attributes is established.The experimental results show that the algorithm greatly optimizes the time efficiency of data exchange.(3)Aiming at the problem of automatic slicing of logging curves and 3D formation drawing,this paper designs and implements a horizon prediction model based on BiLSTM.For the calibrated formation data,the Delaunay triangulation algorithm is used to draw the formation surface to fit the undulating characteristics of the formation.The experimental results show that the accuracy of the prediction result is high.(4)The 3D visualization prototype system of scalar field data such as seismic data,logging data and formation data is designed and implemented.The system realizes the functions of real-time rendering of 10GB geological scalar field data,interactive setting of visual parameters,interactive selection of multi-profile mouse,fusion display of well location information and terrain data,and so on.