海量断层数据的三维重建

Reconstruction of Huge Segment Datasets

对于断层数据的三维曲面重建方法通常是先进行重建,然后对结果进行平滑和简化操作。但是,如果数据量特别巨大,由于受到存储空间的限制,传统方法的执行效率会特别低下甚至无法进行。本文将数据分成连续的若干段分别进行重建,然后再将结果合并,实现了海量断层数据的重建。针对相邻段结果的合并问题,我们设计了一种网格数据存储格式,并基于此提出了相应的合并算法。结果表明,该算法能够很好地保持各段之间的拓扑关系,为后续的网格平滑和简化操作提供了数据基础。

The normal method of reconstruction from segment datasets is extract the isosurface first and then optimize the result. For the huge segment data case, however, the traditional operations have very low efficiency or even can not proceed because of the limit of memory space. In this paper, we divide the original data into some continuous parts and reconstruct each part respectively. The final result will be obtained through merging the mesh data of each part with a special format. The algorithm can keep the topological information used for optimizing the operations very well.