Algorithm Research Based On Dynamic Simulation For Fire And Hair

The tremendous computing workload can embarrass the real-time displayin computer environment. How to simplify the natural scenes with fire and hairis becoming a pivotal technique. Several aspects were probed into as follows:Firstly, this paper goes deeply into the basic theory of particle system.Secondly, according to the appearance of fire, the positions of newparticles are initialized with Gauss random numbers, which are evaluatedapproximately by adding three random numbers. Based on the analysis of themotion of fire, those complex motion fields of fire are simplified to four simplemotion fields. Thus the implementation velocity is enhanced, and real timesimulation becomes possible. At the same time, dynamic motion fields areintroduced. When motion fields move, they make the trajectories of fireparticles time-variant. So the actions of fire particles become more active andthe motion of fire becomes more realistic. Furthermore, we use surfaceparticles instead of dot particles to render fire. Each surface particle cansubstitute many dot particles. So the implementation time is further decreased,and we also use texture map technique to enhance the render quality.Thirdly, a novel approach for combining explicit hair models with clusterhair models was presented.Lastly, the base skeletons were subdivided and grouped to form clusteringhierarchies by using a quad-tree data structure during the precomputation. Atrun time, our algorithm traverses the hierarchy to create continuous lever ofdetails on the fly and chooses both the appropriate discrete and continuous hairlever of details representations based on the motion, the visibility, and theviewing distance of the hair from the viewer.The modeling of fire and hair is improved, which obtained noticeableperformance improvement, with little loss in visual quality.