Geometric-based Special Effect Simulation For Water And Cloud

With the vast demands in computer games, film and TV, education andentertainment, computer animation techniques are used widely. Special effectssimulation is a digital technique which is different from the physical simulation withthe objective of visual reasonability. We propose some practical methods forsimulating the special effect simulation for water and clouds. They can be used indigital entertainment.We develop a special effect simulation framework for water and clouds. Byadopting the spheres and ellipsoids as initial primitives respectively, we propose twofast and automatic surface reconstruction algorithms with spherical- andellipsoidal-blobby model. Based on spherical-blobby implicit surfaces, we present anovel morphing algorithm between 3D models which includes cellular matching andhierarchical matching to create a blob-like special effect. Benefited from theellipsoidal-blobby implicit representation, we propose two animation control schemesfor the motion of blobs to simulate the metamorphosis of clouds naturally——theaggregation scheme and the diffusion scheme. These methods can simulate cloudanimations with shape control.Based on the sphere-tree construction algorithm, we propose an ellipsoid-treeconstruction algorithm which uses ellipsoids as bounding volume primitives insteadof spheres. In our algorithm, an error-based ellipsoidal merge algorithm and avariational ellipsoid optimization algorithm are designed. Experimental resultsdemonstrate that, when the number of bounding volume primitive is small, the samelevel of ellipsoid-tree can segment the model more meaningfully than the sphere-treeapproach with better approximation.We propose a special effect algorithm for simulating water-wave on a modelsurface. In our scheme, we use a damped sinusoidal function based on the geodesicdistance so that the correct simulation can be obtained. During the simulation, themost important problem is to compute the geodesic distance as fast and exactly as possible. We modify the approximating method and make it work for the concave andmultiple sources.Combining with the sketch-based technique and the blobby model implicitrepresentation, we develop a new fast cloud modeling system. In our modeling system,the user sketches a silhouette which is approximated by ellipses, and then the systemcreates a super-ellipsoidal-blobby model. By adopting the fast cloud renderingtechnique, various cloud models with specified shape are obtained. A cloud modelwith complex topology and geometry structure can be created by appending newsketches.