在各种三维锥束重建算法中,近似算法由于数学形式上简单,实现起来容易,而且在锥角比较小的情况下,能够取得较好的重建效果,所以在实际中有着广泛的应用。在各种基于滤波反投影的近似算法中, FDK类型的算法一直是实际应用中的主流。其中P-FDK算法(重新排列投影数据再进行重建)又是FDK算法的一种推广。算法中体素是人为划分的, 考虑在实际重建物体时, 我们所定义的体素与其相邻的体素有一定的比例关系。因此提出P-FDK算法的三维_六邻接的改进方法, 并用采集数据进行重建, 试验结果表明, 新方法重建的图像边缘比传统P-FDK方法清晰, 而且还有抑制噪声的作用。

Among all kinds of reconstruction algorithms for 3D cone-beam CT, approximate algorithms are widely used in practice, because they have simple mathematics expressions, and can be easily implemented. The acquired reconstruction effect is good when cone angle is small. Among various approximate algorithms based in filtered back-projection, Feldkamp-type algorithms are the mainstream methods in practice all the time. The P-FDK(Rearrange the projection data then get reconstruction) is one algorithm of the extending FDK algorithms. To satisfy the voxel of the artificial division and considering the ralation between the defined voxel in real objects and others close to it, an improved technique for P-FDK algorithm was proposed, which is a 3D_6neighborhood P-FDK algorithm. The data used for reconstruction were also from the lab, the practical experiment indicated that the improved method would give a better result in the reconstructed image verge, and also the new method would restrain the noise to a certain extent.