プラズマ計測のためのトモグラフィ解析法

概要

論文の詳細を見る
From the standpoint of data analysis, the computerized tomographic imaging of plasma is reviewed. In starting with the mathematical formulation of the image reconstruction from projections, a variety of numerical methods useful for sparse-data tomography in plasma experiments are taken up and classified into a few categories of techniques for regularizing the ill-conditioned least-squares solution of the linear equation: (1) the series expansion model fitting, (2) the penalty function minimization of Tikhonov-Phillips and nonlinear one of maximum entropy, and (3) modifications with the Fisher information and the Hopfield neural network. Basic concepts of composing plasma images with basis functions a priori chosen and numerically generated, and the filtering of Wiener type and its statistical optimizations of Akaike and Wahba are explained instructively in regarding successful results and by exhibiting undergoing works of plasma imaging in nuclear fusion research.
社団法人プラズマ・核融合学会の論文
2006-07-25