论文部分内容阅读
为了校正窄角扇束工业CT扫描重建的固体火箭发动机图像中的条状伪影,分析了条状伪影的产生原因.针对条状伪影是由重排和数据对齐后固体火箭发动机投影数据的非光滑连续性引起的,提出了条状伪影校正的两种算法.两种算法均以中间平移步数的投影数据为基准进行投影数据的重排,一种将重排和数据对齐过程分开处理;另一种将重排和数据对齐过程合并处理.为了验证两种条状伪影校正算法的有效性,采用窄角扇束工业CT对某型固体火箭发动机进行3种不同扫描步长的扫描和重建.结果表明重排和数据对齐过程分开处理的校正算法只有在小步长扫描的情况下重建效果较好,条状伪影较少;而重排和数据对齐过程合并处理的校正算法在3种不同的扫描步长下图像重建质量均较好,条状伪影较少.
In order to correct the strip artifacts in solid rocket motor images reconstructed by narrow-angle fan beam industrial CT scan, the causes of strip artifacts are analyzed.The strip artifacts are obtained from the rearrangement and data alignment of solid rocket motor projection data Which is caused by the non-smoothness continuity of the image, two algorithms of strip-shaped artifact correction are put forward.All the algorithms calculate the rearrangement of the projection data based on the projection data of the intermediate translation steps, And the other will rearrange and data alignment process combined.In order to verify the validity of the two kinds of strip artefacts correction algorithm, narrow-angle fan beam industrial CT of a solid rocket motor for three different scanning steps The results show that the reconstruction algorithm and the data alignment process separately process the correction algorithm only in the case of small-step scanning reconstruction effect is better, the strip artifacts less; and rearrangement and data alignment process consolidation correction The algorithm has better image reconstruction quality under three different scanning steps and less strip artifacts.