Development of tomography analysis method using sparse modeling

Computed Tomography (CT) is a critical measurement technique for analyzing the spatial structure of various plasma phenomena. However, in plasma and nuclear fusion research, the degrees of freedom for detector placement are often restricted due to the specialized environment, necessitating image reconstruction from a limited number of projections. In future fusion reactors, it is anticipated that detector placement will be further constrained, raising concerns that conventional CT methods may not adequately reconstruct the spatial structure of plasmas. In this study, we developed a tomography method combining Fourier-Bessel series expansion and sparse modeling with L1 regularization to reconstruct images in cases where detector freedom is restricted and there are numerous missing information regions. Sparse modeling proves to be an effective approach for data obtained with limited line-of-sight information. We applied the developed method to reconstruct the ^mode structure in the poloidal cross-section driven by MHD instabilities. The results demonstrated that our method successfully replicated the mode structure assumed in the phantom, even with data significantly constrained by limited lines of sight. (J.P.N.)