【作者】 闫桢楠；

【导师】 顾力栩；

【作者基本信息】 上海交通大学 ， 计算机应用技术， 2009， 硕士

【摘要】 从计算机诞生至今,计算机的应用已经遍布人们生活的方方面面,从科学计算到日常娱乐无不有计算机的影子。计算机的发展加速了不同学科交流融合的步伐,如今数学、物理、生物、医学等都与计算机科学有着密切联系,而医学虚拟手术模拟系统更是将这些学科全部有机地结合在一起。虚拟手术系统作为虚拟现实技术的一种,与人体软组织的三维建模都已经成为是计算机图形学的重要研究课题。而其中的软组织形变计算研究也是实时虚拟手术系统和人机交互系统中最具挑战性的内容。本文讨论了虚拟整形手术系统的需求分析及模块设计,并对其中的软组织形变仿真及其在虚拟手术系统中的应用展开详细分析和实验。考虑到软组织的非线性特性,为了兼顾计算的准确性和计算效率,重点研究讨论了非线性有限元模型的实现及其改进方案。本研究在传统有限元计算的基础上,提出创新性的混合计算方案,结合线性的快速和非线性的准确性计算软组织形变,并实验验证其可行性和效率;同时介绍了三维网格重建算法、图形处理单元的并行计算加速、人机交互技术等相关内容。本文的主要工作和创新点如下:1.详细讨论了有限元方法、相关的力学概念以及对应的数值解法,并针对大形变和材料的非线性问题对求解算法进行整理优化。2.讨论了有限元分析中网格模型的建立与选取。3.本文在几何非线性和材料非线性的混合非线性求解算法中融合了线性近似的计算,提升算法的计算效率。4.利用图形显示卡的并行计算特性进行程序优化。5.建立验证模拟结果的较合理的机制,通过实验验证形变模拟的真实性和有效性。6.根据实际实验结论,选取适当的网格模型,并对算法中的参数和材料参数的设定进行调整优化,以达到计算效率和模拟真实性的最优平衡。7.本论文的新型算法成功地将非线性有限元分析应用于整形手术模拟中,并对形变模拟结果进行应力分析。更多还原

【Abstract】 Since the birth of the computer, it has been applied in every aspact of people’s daily life. Computer has reached areas from scientific computing to the common entertainment. The development of computer speeds up the pace of integration among different disciplines. Mathematics, physics, biology, medicine and computer science, have had close contact now. And medical virtual surgery simulation system is a subject of all these combination.Virtual surgery system, as a kind of virtual Reality technology, and the three-dimensional modeling of human soft tissue have become important research topics in Computer Graphics. The study of soft tissue deformation is also one of the most challenging in real-time virtual surgery and human-computer interaction systems.In this article, we discuss the requirement analysis and module design of virtual surgery system, and propose detailed analysis and experiments on the soft tissue deformation and simulation in virtual surgery system. The soft tissue has non-linear characteristics. So we focus on discussing the non-linear Finite Element Method (FEM) and its improvement in order to achieve a good balance of both accuracy and efficiency of the calculation.In this research, we propose a novel hybrid approach to calculate soft tissue’s deformation. It’s based on classic FEM, combines the fast speed of linear model and accuracy of non-linear model. Experimental results have been given to evaluate the feasibility and efficiency. Besides, it includes the introduction of three-dimensional meshing algorithm, parallel computing by Graphics Processing Unit (GPU) to accelerate, as well as human-computer interaction technology.The main works and innovation points are described as follows:1. The detailed overview of the FEM, related mechanics theories, and numerical algorithm. Optimize the organization of solution for problems with large deformation as well as non-linear material model.2. Discussion on meshing algorithm and model selection in FEM analysis.3. We propose hybrid algorithm considering geometrical and material non-linear. And with the proper linear approximation, we could increase the efficiency of FEM solution.4. With the help of GPU parallel computing, we optimize the speed of calculation.5. We propose a rational mechanism to evaluate the simulation results. And validate the authenticity and feasibility.6. Select and optimize mesh model, program coefficient and material parameters based on experiments to find a excellent balance between efficiency and accuracy.7. Our proposed novel algorithm is successfully applied for craniofacial plastic surgery simulation. Also process stress analysis for the simulation outcome.更多还原