Abstract: The Algebraic Reconstruction Technique is a classical image reconstruction method, it is suit for reconstruction from incomplete projection data. In order to improve the reconstruction speed, the relaxation factor is often simplified to 1 or 0, thus the reconstructions usually suffer from salt and pepper noise. A method for improving reconstruction quality was proposed, the reconstructed image was processed by selectively smoothing after each iteration, then the smoothed image was used as the initial image for next iteration. This method is characterized that it incorporates image processing into image reconstruction. Simulation experiments shows that the method is very effective, not only the reconstruction quality is improved, but also the fuzzy phenomena which caused by smoothing using Box template is overcomed

Key words: algebraic reconstruction technique, image reconstruction, incomplete projection data, reconstruction quality

摘要： ART算法是一种经典的图像重建方法，适合于不完全投影数据重建。为了提高重建速度，通常对权因子进行简化，其结果是在重建图像中普遍存在椒盐噪声。提出了一种改善重建质量的方法，在每次迭代后对重建图像进行有选择的平滑，将平滑结果作为下一次迭代的初值，其特点是将图像处理和图像重建相结合。仿真实验表明该方法非常有效，不但提高了重建质量，而且克服了利用Box模板进行平滑所造成的图像模糊现象。

关键词: 重建质量, ART算法, 图像重建, 不完全投影数据