Open AccessAutomatic analysis of double coronal mass ejections from coronagraph images
Author(s) -
Jacobs Matthew,
Chang LinChing,
Pulkkinen Antti,
Romano Michelangelo
Publication year - 2015
Publication title -
space weather
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.254
H-Index - 56
ISSN - 1542-7390
DOI - 10.1002/2015sw001260
Subject(s) - coronagraph , coronal mass ejection , space weather , thresholding , event (particle physics) , computer science , artificial intelligence , physics , image (mathematics) , computer vision , solar wind , astronomy , astrophysics , stars , exoplanet , quantum mechanics , magnetic field
Coronal mass ejections (CMEs) can have major impacts on man‐made technology and humans, both in space and on Earth. These impacts have created a high interest in the study of CMEs in an effort to detect and track events and forecast the CME arrival time to provide time for proper mitigation. A robust automatic real‐time CME processing pipeline is greatly desired to avoid laborious and subjective manual processing. Automatic methods have been proposed to segment CMEs from coronagraph images and estimate CME parameters such as their heliocentric location and velocity. However, existing methods suffered from several shortcomings such as the use of hard thresholding and an inability to handle two or more CMEs occurring within the same coronagraph image. Double‐CME analysis is a necessity for forecasting the many CME events that occur within short time frames. Robust forecasts for all CME events are required to fully understand space weather impacts. This paper presents a new method to segment CME masses and pattern recognition approaches to differentiate two CMEs in a single coronagraph image. The proposed method is validated on a data set of 30 halo CMEs, with results showing comparable ability in transient arrival time prediction accuracy and the new ability to automatically predict the arrival time of a double‐CME event. The proposed method is the first automatic method to successfully calculate CME parameters from double‐CME events, making this automatic method applicable to a wider range of CME events.