A novel inversion algorithm for optical tomography of plasmas

Summary form only given. In plasma emission and absorption spectroscopy, there is a strong demand to reconstruct the plasma distribution from the side-on intensity measurements. While Abel inversion is commonly used for axially symmetric plasmas, tomographic reconstruction techniques are necessary for asymmetric plasmas. Filtered back projection algorithm is mostly used for medical X-ray imaging. However, model function fitting, algebraic reconstruction technique (ART), or Tikhonov regularization are more appropriate for optical diagnostic techniques, where optical access to source is limited by the vacuum chamber viewports. This is an iterative technique which utilizes an a least square fit of exponential functions to the intersection of the measurement line-of-sights. In each iteration, the emissivities at the intersection points are adjusted to reduce the difference between the line integrated experimental measurements and the line integrated theoretical values obtained by the previous least squared exponential curve fit. Additional conditions placed on the plasma distribution were that the emissivities have positive values in the plasma and approach zero at the boundaries. An observation geometry comprised of two perpendicular fans was investigated to simulate the experimental setup.