Premium
Ionospheric‐thermospheric UV tomography: 1. Image space reconstruction algorithms
Author(s) -
Dymond K. F.,
Budzien S. A.,
Hei M. A.
Publication year - 2017
Publication title -
radio science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.371
H-Index - 84
eISSN - 1944-799X
pISSN - 0048-6604
DOI - 10.1002/2015rs005869
Subject(s) - algorithm , algebraic reconstruction technique , tomography , conjugate gradient method , mathematics , iterative reconstruction , regularization (linguistics) , multiplicative function , computer science , artificial intelligence , mathematical analysis , physics , optics
We present and discuss two algorithms of the class known as Image Space Reconstruction Algorithms (ISRAs) that we are applying to the solution of large‐scale ionospheric tomography problems. ISRAs have several desirable features that make them useful for ionospheric tomography. In addition to producing nonnegative solutions, ISRAs are amenable to sparse‐matrix formulations and are fast, stable, and robust. We present the results of our studies of two types of ISRA: the Least Squares Positive Definite and the Richardson‐Lucy algorithms. We compare their performance to the Multiplicative Algebraic Reconstruction and Conjugate Gradient Least Squares algorithms. We then discuss the use of regularization in these algorithms and present our new approach based on regularization to a partial differential equation.