Research On 3D Cloud Visualization Algorithm Based On Multiresolution Method

In the visual simulation study of natural landscapes,cloud visualization is a research hotspot and research difficulty in the field of meteorology.The generation and dissipation of natural clouds are caused by energy conversion and changes in atmospheric physical states,and cloud changes have an important interpretation of the atmospheric change process,so its importance can be imagined in the weather system.Cloud visualization has been widely used in meteorological services,outdoor scene simulation,film special effects,etc.At the same time,cloud as an important weather element is inseparable from weather changes.Drawing a rendering cloud based on real weather data can assist the meteorologist’s forecasting and forecasting work while transmitting real weather conditions.Therefore,in recent years,the visualization of real clouds based on meteorological data has received a lot of attention.At present,the research of 3D cloud mainly focuses on data preprocessing compression and drawing.This paper is based on large-scale massive data multi-resolution 3D cloud visualization,respectively,to improve the pre-processing compression and rendering methods of the simulated cloud,and improve the rendering speed while ensuring the rendering quality.The main research contents are as follows:(1)In terms of cloud pre-processing compression,the graphics hardware storage space caused by large-scale massive data in 3D cloud simulation is limited,and real-time problem cannot be achieved.A unified data compression method based on coefficient of variation is proposed.The method divides the importance of data blocks according to their degree of data turmoil,retains important data,and appropriately reduces unimportant data.The Haar wavelet transform and the classification vector quantization method are used to perform vector quantization and compression on the divided data blocks,and then the rendering strategy is performed by decompressing and rendering in the GPU,which improves the rendering efficiency while ensuring the image fidelity.(2)In the aspect of cloud rendering algorithm,an improved adaptive ray casting algorithm is proposed for the traditional ray casting algorithm,which is slow in drawing and limited in memory.The method designs the sampling frequency according to the characteristics of line of sight and block importance to achieve adaptive sampling.When classifying using the two-dimensional transfer function,the block importance weighting Shannon entropy is set instead of the gradient value of the conventional method as the second dimensional domain of the two-dimensional transfer function to calculate the opacity value and the color value.The experimental results show that the simulation effect of the cloud visualization is very good,the details are obvious,and the drawing speed is obviously improved.