中文摘要：

目的讨论腹主动脉瘤断层图像的分割方法及其网格剖分技术，建立可用于有限元计算的二维数值分析模型。方法采用完全基于形态学的方法完成图像中各个部分的分割，针对每个分割得到的闭合曲线，计算其符号距离函数，然后根据各个曲线的集合关系得到一个最终符号距离函数，在这个距离函数和一个平衡关系的控制下不断使用Delaunay算法，当满足平衡关系或者达到设定的条件时，网格划分终止，有限元模型建立完成。结果实现血管内腔的自动化分割以及血管壁、钙化点等的半自动化分割；对血管内的不同成分划分网格，并可以控制网格的类型和密度；建立血栓与血管壁耦合以及血栓与血管壁、钙化点耦合的两种有限元模型，并进行相关的应力分析。结论分割过程中不需要复杂的计算及提供初始化曲线，分割快速、准确；网格划分算法可以产生高质量的网格，网格的生成易于控制；产生的网格可用于实际的计算。

英文摘要：

Objective To develop an automatic segmentation and mesh generation technique for abdominal aortic aneurysm （AAA） and to build a 2D numerical analysis model that can be used for finite element analysis （FEA）. Methods A method that totally based on morphology processing was developed to segment all the components of the AAA. For each closed curve that obtained, its signed distance function was then calculated. According to the set relationships between each curve, the final signed distance function was calculated. Under the control of this function and an equilibrium relationship, iterated Delaunay algorithms were used until the equilibrium relation- ship was satisfied or the set conditions were reached. Then the program ended and the finite element model was generated. Results Automatic segmentation of the lumen as well as semiautomatic segmentation of the wall and calcification were achieved. Different parts of the AAA were meshed, and the type and density of the mesh could be controlled. Two finite element models were established for stress analysis： one was the coupling mesh of the thrombus and the wall, and the other was the coupling mesh of the thrombus, the wall and the calcifications. Conclusions An automatic segmentation and mesh generation algorithm with high accuracy has been developed, without any complicated computation or initial curve. The mesh generation algorithm tends to produce high quality meshes and the generation is easy to be controlled by only two parameters. The generated mesh has been veri- fied to be useful in FEA simulation.