To avoid grid reconstruction over point-based model and synthesize the sample texture to pointed set model directly, an algnrithm of 3D texture synthesis based on point set model is researched. Firstly, the neighborhood relation among the discrete sampling points is built by KDtree. Then mapping relation between each point to be synthesized texture and the sample image is gotten by projecting and setting up the matching window. According to the mapping relation the best matching point with texture value is found in the input sample image. Finally, the value is assigned to the point to be synthesized and meanwhile multi-point synchronous assignment is achieved via the neighborhood relation. The experimental results show that the new technique can produce smooth texture synthesis as well as keeping the continuity and flatness of visual effects. In contrast with traditional 3D texture synthesis, this algorithm uses less time and is more flexible and more controllable and of robustness.%为了避免网格重建过程而直接在三维点采样表面上进行纹理合成,研究了在点模型上的三维纹理合成的算法.利用KDtree这种数据结构在离散采样点上构建点与点的邻域关系,通过投影和建立匹配窗得到每一个待合成纹理的点与样图的映射关系,根据映射关系在样图中找出与待定点纹理值最佳匹配的点.最后将该匹配点的纹理值赋给待定点,同时结合邻城关系实现多点同时赋值.实验结果表明,该算法能有效地保持了生成的纹理结构和视觉效果的连续性和平滑性,具有耗时短、灵活可控、较好的鲁棒性等特点.
展开▼
