TomograPy: A Fast, Instrument-Independent, Solar Tomography Software

TL;DR

TomograPy is an open-source, fast, and scalable software for solar tomography that handles multi-source data and models the temporal evolution of solar structures, enabling improved 3D reconstructions.

Contribution

It introduces a generic, efficient, and multi-source compatible software package for solar tomography with parallelized algorithms and flexible physical models.

Findings

Demonstrated multi-spacecraft tomographic inversions with STEREO data

Achieved linear scaling with data volume and computational cores

Supported static and dynamic solar structure modeling

Abstract

Solar tomography has progressed rapidly in recent years thanks to the development of robust algorithms and the availability of more powerful computers. It can today provide crucial insights in solving issues related to the line-of-sight integration present in the data of solar imagers and coronagraphs. However, there remain challenges such as the increase of the available volume of data, the handling of the temporal evolution of the observed structures, and the heterogeneity of the data in multi-spacecraft studies. We present a generic software package that can perform fast tomographic inversions that scales linearly with the number of measurements, linearly with the length of the reconstruction cube (and not the number of voxels) and linearly with the number of cores and can use data from different sources and with a variety of physical models: TomograPy (http://nbarbey.github.com/TomograPy/), an open-source software freely available on the Python Package Index. For performance, TomograPy uses a parallelized-projection algorithm. It relies on the World Coordinate System standard to manage various data sources. A variety of inversion algorithms are provided to perform the tomographic-map estimation. A test suite is provided along with the code to ensure software quality. Since it makes use of the Siddon algorithm it is restricted to rectangular parallelepiped voxels but the spherical geometry of the corona can be handled through proper use of priors. We describe the main features of the code and show three practical examples of multi-spacecraft tomographic inversions using STEREO/EUVI and STEREO/COR1 data. Static and smoothly varying temporal evolution models are presented.