Improving the efficiency of small-angle x-ray scattering computed tomography using the OSEM algorithm.

Small-angle x-ray scattering computed tomography (SAXS-CT) is a nondestructive method for the nanostructure analysis of heterogeneous materials. However, the limits of a long data acquisition time and vast amounts of data prevent SAXS-CT from becoming a routine experimental method in the applications of synchrotron radiation. In this study, the ordered subsets expectation maximization (OSEM) algorithm is introduced to improve the efficiency of SAXS-CT. To demonstrate the practicability of this method, a systematic simulation and experiments were carried out. The simulation results on a numerical phantom show that the OSEM-based SAXS-CT can effectively eliminate streaking artifacts and improve the efficiency of data acquisition by at least 3 times compared with the filter backprojection algorithm. By compromising the reconstruction speed and image quality, the optimal reconstruction parameters are also given for the image reconstruction in the OSEM-based SAXS-CT experiments. An experiment on a bamboo sample verified the validity of the proposed method with limited projection data. A further experiment on polyethylene demonstrated that the OSEM-based SAXS-CT is able to reveal the local nanoscale information about the crystalline structure and distributional difference inside the sample. In conclusion, the OSEM-based SAXS-CT can significantly improve experimental efficiency, which may promote SAXS-CT becoming a conventional method.