基于图像空间的三维动画线绘制方法研究-计算机科学与技术专业论文.docxVIP

优秀毕业论文 精品参考文献资料 AbstractAbstract Abstract Abstract With the development of computer graphics in recent years，non-photorealistic rendering(NPR)has been paid more and more attention．As the most concise and communicative technique of the NPR，line drawings Can be remarkably efficient at conveying the information of objects as much as possible witll a minimum of visual distraction．Line drawings is also the basic element of other NPR techniques，it can be combined with other methods in order to produce different kinds of rendering effect． There are two main categories of line drawings with three—dimensional models， object method and image space method．The object space method works directly on the mesh of the object itself．This method can get line drawings谢t11 high quality， however the efficiency of the method depends on how complicated the object mesh is． While compared with object method，image space method is much more easy to implement and also much more efficient，but the quality of lines is less desirable．This is because the quality of lines in image space is closely related to the quality of the intermediate rendered image，it’S very sensitive to the noise and trivial details appeared in the rendered image．As for line drawings of 3d animations，the surface normal of the mesh needs to be recomputed every frame，the computation of which is both the critical step and bottle neck of line drawings．Although we can utilize the existing methods to generate line drawings frame by frame，it cannot be suited for large scale data． This dissertation focus on the coherence between adjacent frames in animated models，based on the image space method and make a improvement on it．We proposed a line drawing framework which Can be used for large scale skinned meshes． This method exploit the efficiency of image space method and also make a progress on the quality of lines．The basic idea is follows．First of all，we speed up the computation of surface normal by utilizing the coherence between fra