虚拟内窥镜系统中的自动导航

目的 研究基于PC机的虚拟内窥系统中的自动导航方法。方法 首先设计了一种基于位编码和坐标变换的快速3D细化算法来提高提取器官骨架的运算速度；然后采用2次跟踪技术裁剪骨架分支得到所需的导航路径；最后提出一种3遍3次Bezier平滑导航路径算法和Bezier插值算法，使导航路径至少能达到C1连续且较为平坦，并以平滑的导航路径上的点控制虚拟摄像机位置，以对应点处的切线(一阶导数)控制虚拟摄像机方向。结果 对经过分割后获得的人体气管体数据进行自动导航实验，结果表明，在自动导航中虚拟摄像机能保持远离气管内壁，而且相邻帧之间过渡比较平滑，能产生效果很好的自动导航动画。结论 方法实现简单。基本上能满足虚拟内窥镜系统临床应用的要求．

Automatic Navigation for Virtual Endoscope

Objective To explore the method of automatic navigation for PC based virtual endoscope system. Method A 3D thinning algorithm based on bit encoding and coordinates transform was designed to speed up the calculation of extracting the skeleton. The small branches of the skeleton were trimmed using two pass tracking technique. Then, the central line was smoothed by using tri pass cubic Bezier smoothing method. A plainer central line was obtained by using aforementioned algorithms and the resulted line had C 1 continuity at least. The virtual camera would move along this smoothed navigation path point by point and the camera direction followed the tangent of the central line. Result Automatic navigation experiments were carried out with the volume data of segmented human trachea and the results showed that the camera was kept off the inner wall of the trachea during the automatic navigation and the animation transition was gentle and smooth. Conclusion The proposed algorithms are simple to realize and meet the requirements of virtual endoscope in clinic.