采用图形加速的三角网格实时切分

摘要

提出一种采用图形加速的三角网格模型实时切分的方法。针对传统的三角网格实时切分方法普遍效率不高的问题，提出利用 OpenGL 的拾取机制的快速、有效，将屏幕曲线映射到模型上，得到切分边缘的三角面片。并利用网格的 AIF（Adjacency and Incidence Framework）数据结构和当前图像场景的视角矩阵优化网格模型交线生成追踪过程。然后将相交的三角面片重新三角化，构建新的拓扑结构。最后分离模型，实现模型的快速切分。实验结果表明，该方法能够快速有效地完成模型的实时切分。%The paper proposes a method which uses graphics acceleration to cut triangular mesh in real-time.For the problem of generally inefficient in traditional triangular mesh real-time cutting method,we propose to use the picking-up mechanics in OpenGL to map the screen curve onto model fast and effectively,and through the use of AIF (adjacency and incidence framework)data structure and the perspective ma-trix of current image scene to optimise the tracking process of the generation of mesh model intersecting lines.Then we re-triangulate the inter-secting triangular facets,and construct the new topology structure.Finally,we split the model and implement the fast cutting of model.Experi-mental results show that this method can accomplish real-time triangular mesh model cutting fast and effectively.