Study On The Natural Morphing Technologies Of Object Space

As one of the key techniques for computer animation, morphing technique has attracted much attention in research recently. Despite of computer animation, morphing technique has also been widely applied to areas such as industrial modeling, virtual reality, science computation visualization, film stunt and so on.There are two main branches of morphing research, one is based on object space, and the other is based on images. The morphing of planar polygon is an important component of object morphing, which is also a hotspot. A lot of approaches have been proposed for morphing planar polygon, but most of them have a certain limits and hard computational loads. Therefore, this paper makes researches on two-dimensional morphing of object space, in order to improve the applicability and to get more natural morphing result while reducing the computational work needed for metamorphosis.In the morphing process, how to make edge self-intersection-free and how to preserve the geometrical properties of the intermediate states are two challenges. Surazhsky and his colleagues first gave their guaranteed intersection-free polygon morphing approach based on convex combination. However, to get the compatible triangulations in a common convex boundary is time-consuming, especially when the shapes of source and target object are quite different, and usually the morphing result is not satisfied. Although the convexity-preserving metamorphosis method proposed later has further reduced the number of Steiner points needed for compatible triangulation, the computational work needed for metamorphosis is still hard. In order to get the benefit of intersection-free polygon morphing method furthest, the morphing approach based on compatible triangulation is determined to be adopted. Researches both for reducing the computational work and improving the quality of morphing have been done in this paper.In order to reduce the number of Steiner points, an improved similarly compatible triangulation method is proposed, and the compatible triangulation method based on polygon convex decomposition is improved. The improvements for both the computation of visible polygon for given point in a simple polygon and the minimum link distance path between given vertex pair in a simple polygon have been made, which are used in the fast polygon decomposition algorithm based on point visibility. A hybrid morphing approach has been proposed to reduce the computation load and get more natural morphing effects.A polygonal approximation approach based on Genetic Algorithm (GA) is proposed, whose aim is to preserve the visual features of contours of source and target objects, and help for getting more natural morphing effects. By appointing the starting correspondent vertex pair for solving the correspondence problem, the applicability of the morphing method has been strengthened. A new convex hull algorithm and a new kernel algorithm have been proposed to speed up the computation of the convex hull and simplify the adjustment to locations of the newly added Steiner points on boundaries of the magnified convex hulls respectively. In order to improve the quality of morphing, a geometrically guided morphing approach has been presented, which preserves the geometrical properties of the intermediate polygon more uniformly than controllable convex combination morphing and intrinsic morphing approaches proposed by Surazhsky and his partners. Through experiments, it is verified that the hybrid morphing approach based on similarly compatible triangulation not only has low computational complexity, but also could get natural morphing effects even if the shapes of source and target object are quite different. The stronger applicability and higher efficiency of the hybrid morphing approach have made a massive basis for further researches on natural morphing of object space.