Reconstrução 3D de imagens tomográficas de raios-x de arcos coronais em ambiente paralelo

Abstract

Space researches have been carried with the objective to forecast solar explosions, through images obtained by satellites equipped with X-ray telescopes. The highresolution solar images allow the study of the spatio-temporal dynamics of plasmamagnetic structures in arc shape, denominated coronal loops. The coronal loops can provide important information about the solar explosions dynamics that may cause serious perturbations in terrestrial communication systems. However, this information is subject to errors due to the spatial limitation imposed by bidimensional images. Motivated by the need to extract information of the tridimensional structure of the coronal loops, a method for 3D reconstruction of coronal loops from two X-ray tomographic images captured by the Japanese satellite Yohkoh was developed. This method of 3D reconstruction based on image morphing generates the intermediate images of the transversal sections of the loop using a image warping method controlled by a Bezier curve and an approach to Image 3D Interpolation using Bayesian Estimation Theory. Due to the need for the 3D reconstruction of the coronal loops in a reasonable time for study of the spatiotemporal dynamics, a parallel program was implemented to execute in a high performance parallel machine.