Research On Optimization Method For Volume Rendering Of Large-scale Scientific Data

Scientific visualization has made great progress in recent years.For data analysis of scientific visualization,volume rendering is one of the most widely used methods.Volume rendering can represent obscure data structure with intuitive 3D images,and analyze the internal and hidden information of data through interactive methods.However,as the increasing of scientific data sets day by day,such huge scientific data cannot be taken to complete data pre-processing,volume rendering,and real-time interaction efficiently on a single computer.For the problem,this paper mainly researches volume rendering optimization methods for large-scale scientific data from two aspects:(1)large-scale scientific data volume rendering based on WebGL;(2)large-scale scientific data volume rendering based on parallel architecture.For WebGL,it preprocesses the original scientific data and the browser loads it into GPU to take Ray-casting algorithm.In order to further reduce the memory usage and drawing time,this paper uses a dimensional compression method,which greatly reduces the space occupation of the original data.Parallel architecture volume rendering is to divide original data into blocks and to trasfer them into a cluster of multiple computers connected by network for taking volume rendering using VTK.This paper uses the Isabel Hurricane time-varying data set to design the algorithm for the above two aspects and conduct experimental verification.Besides,this paper designs a volume rendering system based on the B/S architecture,data preprocessing,volume rendering,and real-time Interactions are integrated into the system,and the whole volume rendering process for large-scale scientific data is realized.Testing result show that the algorithm in this paper can run effectively.In the volume rendering system based on parallel architecture,By comparing with the rendering in single machine,the rendering time can be greatly reduced.In the WebGL-based volume rendering system,the frame rate can reach more than 50 fps,and through dimensional compression method,the evolution of the hurricane in real time can be observed in a browser.