Large-scale Cloud Simulation Based On Real Weather Data

Cloud is a natural phenomenon and a common sight in the sky.3D cloud modeling and rendering research has always been a key part of computer graphics.Using computer to generate cloud can be applied to many areas.Adding 3D clouds to the sky can make the outdoor scenes visually,and often used for flight simulation and outdoor battlefield simulation.In special effects movies,3D video games,and the increasingly popular virtual reality field in recent years,3D clouds also play an important role.In addition,using meteorological data to generate 3D cloud can help meteorologists to observe weather phenomena more intuitively,and play an important role in weather forecasting and artificial weathering.In recent years,due to the improvement of computer hardware performance,the speed of data processing has been greatly improved,and the means of obtaining meteorological information has become more and more abundant.3D cloud simulation based on meteorological data information has gradually become the mainstream.Clouds generated by this method not only have realistic visual effects,but also have physical meaning.This thesis conducts large-scale 3D cloud simulation research based on real meteorological data,studies the application method of meteorological data in cloud simulation technology,and proposes a new adaptive grid structure for cloud modeling.By analyzing of the illumination model inside of the clouds,proposes a new high-order anisotropic multiple scattering model for cloud and applies it to cloud rendering.The specific research contents are as follows:In the aspect of cloud modeling algorithm,in order to solve the problems of high modeling complexity and low rendering efficiency for the existing real weather forecast cloud data simulation methods,a three-dimensional visualization method of weather forecast cloud data based on the adaptive far-field grid structure for region of interest is proposed.The method first extracts the extinction coefficient of the cloud as a cloud simulation factor,then extracts the region of interest to generate an adaptive far-field grid structure and applies it to the modeling of cloud particles,optimizing the number of particles of other regions while maintaining the fine resolution of the region of interest,and finally rendering the 3D cloud image.The simulation results based on the WRF model weather forecast cloud data show that the grid structure can speed up the rendering speed based on the guaranteed rendering quality,and can better display the shape and structural characteristics of the real cloud.In the aspect of cloud rendering algorithm,the existing illumination model mostly ignores the high-order scattering and causes the lack of realism of the generated cloud image.A cloud rendering method based on high-order anisotropic multiple scattering model is proposed.This method expands the idea of the most probable path algorithm,treats the cloud as a parallel plane,proposes the concept of heat collecting region,contains most of the incident ray that finally reaches the viewpoint,and uses the normal transfer function of light to effectively simulate the high order.scattering.And improved the ray casting rendering algorithm,using acceleration step sampling and composite interpolation method,instead of the uniform sampling and trilinear interpolation method in the traditional ray casting method.The simulation results based on WRF model meteorological cloud data show that the rendering method can effectively improve the realism of the lighting effect of generating cloud images,and has higher computational efficiency when drawing.