Simultaneous Coronal Hole Segmentation from Solar EUV Images using the Image Co-segmentation Technique

Abstract

Accurate detection of coronal holes (CHs) is an important issue in the domain of solar physics, since, CHs plays an vital role in controlling the overall space weather and generation of solar wind. In this paper a novel application of image co-segmentation (Co-seg) has been proposed for accurate detection of coronal hole simultaneously in two solar disk images of different band. For the task of Co-seg circular Hough transform inspired active contour method (ACM) have been proposed. Here the circular Hough transform has been introduced to setup the initial contours for the ACM based Co-seg. In later stage, the initiated contours has been evolved based on the energy function by incorporating the color features. At the final stage, when the contours do not deform any more is considered as the final contours position and the regions inside the contours is extracted as the CHs regions from both the images under consideration. The experiment has been conducted on few benchmark datasets and the corresponding results have been compared with the results of other existing techniques. The compared outcomes validate the performance of the applied technique in detection of coronal holes in solar disk. © 2020 URSI.