【作者】 关防利；

【导师】 潘荣江；

【作者基本信息】 山东大学 ， 计算机软件与理论， 2013， 硕士

【摘要】 基于体数据的三维织物仿真,是国内外计算机图形学界一个崭新的研究领域。体绘制就是依据三维体数据把获取的所有层次细节同时展现在二维图像上。采用体数掘进行织物仿真不仅能够反映出实际纱线表面复杂的捻度和毛羽效果,还能表现出纱线表面纤维的互相交叉、遮挡和半透明特性,而这些效果和特性是二维图像技术和三维面模型技术很难实现的。论文介绍了国内外几种常见的织物仿真方法并进行了对比。首先提出了一种纱线体数据模型建立的新方法,然后建立了织物体数据模型,最后根据弹簧-质点模型实现了织物体模型的运动仿真。论文的主要研究内容及贡献包括以下几个方面：首先,建立单根纱线的体数据模型。组成体数据的每一个数据单位称之为体素。在本文中每个体素由R、 G、B、A四个数值组成,由显微CT扫描的纱线切片只可以获取纱线的密度信息即A值,而没有颜色信息。为了给纱线体数据模型赋上颜色,本文使用平板扫描仪扫描真实纱线,得到其彩色图像,采用一个迭代算法使显微CT扫描体数据投影图片的RGB值与平板扫描仪扫描纱线图像的RGB值的统计均值一致即可。然后,建立整个织物的体数据模型。织物由多根纱线交织而成,纱线的中轴线采用三次B样条曲线来表示,根据织物结构中经纬线的中轴线位置信息对纱线体数据进行变换和填充,得到织物的体数据模型,利用体绘制方法进行绘制,得到真实感较强的仿真效果,可表现出纱线的毛羽、捻度等效果。最后,实现织物体数据模型的运动仿真。以质点-弹簧模型为基础,结合经典的牛顿力学模型和运动学定律分析织物模型的受力情况,建立基于体数据织物模型的运动方程,给出了体数据模型中每个质点的位移量。每经过一个时间步长,更新一次中轴线的位置信息,再重建织物体数据模型,实现运动效果。利用C++和OSG对上述方法进行了实验,实验结果表明基于体数据的织物仿真结果更加真实自然。更多还原

【Abstract】 Simulation of Jacquard Fabrics based on volume data is a new direction of research in computer graphics field in recent years. The volume rendering technique is to show all the details at the same time on a two-dimensional image based on volume data. With this technology, it can reflect not only the complex twist and hairiness effect of yarn surface but also fibers crossing and translucent properties, which can not be achieved by other simulation methods such as two-dimensional image technology and three-dimensional surface modeling techniques.The paper introduces several common fabric simulation methods and gives their comparisons. First of all, a new method is proposed to build volume data model of single yarn. Then a volume data model of fabric is established. Finally we simulate dynamic fabrics with volume data model based on the principles of mechanics.The main research content and contribution includes the following aspects:Firstly, a volume data model of single yarn is established. The unit data which composed of the whole volume data is called voxel. The four components in each voxel are R, G, B, A. But the yarn slice obtained by the micro-CT scan contains only A and does not contain R, G, and B. In order to obtain a color volume data model of yarn we use a flatbed scanner to scan real yarn. An iterative algorithm is being adopted to make the two statistics mean consistent.Secondly, the volume data model of fabric is established. Multiple yarns interweave in fabric, and the central axis of yarn is represented by cubic B-spline curve. The final fabric volume model is refined by interpolating warped yarn data according to axis position of warp and weft yarns. The effects of hairiness and twist of yarns are manifested via volume rendering.Finally, dynamic simulation of fabric volume model is achieved. Using the classic Newtonian mechanics model and the law of motion, we analyze the situation of the model based on the mass-spring model, and gives the composition and expression of internal and external forces of each particle in the volume model. The fabric volume model is dynamically simulated according to the position information of the central axis which is updated after each time step.We implemented the algorithm using C++and OSG. Experimental results show that the model is feasible and the simulation results are natural.更多还原