基于人体结构断层图像的三维建模与网格剖分优化

3-D modeling and mesh optimization using cross-sectional images

建立了基于断层图像的近似于人体真实结构的三维表面模型,并根据曲率优化网格,优化了有限元计算的前处理,以利于临床力学研究。基于Delaunay三角剖分等相关理论,从临床计算机断层、磁共振和可视人切片等二维断层图像提取轮廓散点,按照层结构有序排列的特点,重构三维表面模型。再根据人体真实结构表面曲率的不同,调节和控制网格精度。曲率大的地方细化网格,曲率小的地方稀疏网格。实现了多分辨率建模和自适应剖分。

A 3-D surface model which approximates the real human body was reconstructed using cross-sectional images for clinical biomechanical studies. Surface meshes were optimized from the surface curvature. Contours in the 2-D slices were extracted from clinical CT, MR, and visible human slices. A restricted 3-D surface modeling method was used to reconstruct 3-D human bone models using Delaunay triangulation. The mesh precision was optimized based on the surface curvatures on body. Smaller elements were used with large curvatures and large elements with smaller curvatures. The method ensures multi-resolution modeling and self-adaptive mesh generation and automates the pre-processing for finite element analysis.