Abstract Title: *
A technique for automatic image sequences alignment

Authors: *
Dimitar Garnevski, Petya Pavlova, Nikola Petrov

Institution: *
TU Sofia, branch Plovdiv

Preferred format: *
poster presentation

Abstract Text: *
Many methods for processing a series of images using algorithms for correlating two images
rely on the initial alignment of the objects on the input frames. Tracking the processes in the
solar corona requires exact alignment in the positioning of the sun in the analyzed images. The
question remains of the greatest importance when the images are obtained using terrestrial
instruments - a coronograph. This is mainly due to the fact that manual positioning of the
coronagraphs can lead to a relatively large deviation in the position of the objects in two
consecutive frames. Which in turn will require manual image alignment during processing or
analysis? The main task of this study is to create a technique for arranging two images of the
solar corona that are obtained by a coronograph. The images consist of two parts - part of the
core area of the crown and a black area of artificial moon. Automation is based on the position
of centers of circles that approximate the gradient of the light distribution. Progress in solar
corona registration technology has resulted in improved spatial and time-resolved data. Hence,
an increase in the amount of information to be processed, both by increasing the resolution of
the images and by reducing the intervals between the records of two consecutive images. That's
why we target the implementation of the algorithm by applying advanced image processing
technologies, including parallel computing.