基于先验图像约束压缩感知多能CT重建算法

Prior Image Constrained Compressed Sensing CT Reconstruction Based on Multi-Energy

由不同材质构成且材质之间密度相差较大的工件进行单一能量CT重建时无法获取完整内部结构.为获得结构信息完整的高质量CT重建图像,研究了基于先验图像压缩感知多能重建方法.首先从低到高依次采集多个能量下的投影数据,并用凸集投影-全变分最小化(POCS-TVM)算法对最低能量的投影数据进行CT重建,然后,将重建好的图像作为先验信息,利用先验图像压缩感知算法(PICCS)对下一组能量下的投影进行重建,重建后的图像再次作为新的先验信息重复以上步骤,依此类推直到最高能量的数据重建,以此达到完整重建.结果表明该方法可以有效减少因投影缺失而导致的伪影并保护低密度边缘.

When the workpiece was irradiated by a single energy,it is difficult to obtain the complete CT image of workpieces which were composed of different materials with large density difference.In order to accurately reconstruct CT images of workpieces and obtain a high-quality CT image,the prior image constrained compressed sensing CT reconstruction based on multi-energy has been studied.Firstly,projection data were collected from low energy to high energy in turn and POCS-TVM algorithm was utilized to reconstruct the image at the lowest energy.Then,the reconstructed image at the lowest energy was regarded as prior information and utilizing the prior image constrained compressed sensing algorithm reconstructs next energy CT image.The reconstructed image was used as the prior information again.Repeat the above steps until the highest energy.The simulation results show that the method can effectively reduce the artifacts due to under-sampling projection.And the low density edges are well protected.