三维重建中形状插值的快速算法研究

Shape-Based Interpolation Algorithm for 3D Reconstruction

提出一种用于三维重建的层间数据插值算法 该算法首先利用数学形态学算子提取层内图像骨架 ;然后利用骨架匹配算法获取层间骨架点集的平移、旋转和缩放信息 ,并利用此信息的线性插值获得插值层图像的骨架点集 ;最后根据形态骨架的重建算法获得插值层图像 文中算法只对骨架点集进行操作 ,具有较小的运算复杂度 ,较好地保留了原始图像的形状信息

First, it uses morphological operators to extract the skeletons for adjacent tomographic image pair Then, skeleton matching algorithm is employed to obtain the information of translation, rotation, and scaling between skeleton pair, which is applied to get the skeleton of any interpolated image between the tomographic image pair by linear interpolation Finally, the interpolated images are reconstructed from their skeletons with morphological operators Since the new algorithm only processes the skeletons, it has less computational complexity and preserves the shape information of the original object to some extent Experimental results illustrate the effectiveness of the proposed approach