Coronal Hole Analysis and Prediction using Computer Vision and LSTM Neural Network 后印本

摘要: As humanity has begun to explore space, the significance of space weather has become apparent. It has been established that coronal holes, a type of space weather phenomenon, can impact the operation of aircraft and satellites. The coronal hole is an area on the sun characterized by open magnetic field lines and relatively low temperatures, which result in the emission of the solar wind at higher than average rates. In this study, To prepare for the impact of coronal holes on the Earth, we use computer vision to detect the coronal hole region and calculate its size based on images from the Solar Dynamics Observatory (SDO). We compare the coronal holes for each region of the Sun and analyze the correlation. We then implement deep learning techniques, specifically the Long Short-Term Memory (LSTM) method, to analyze trends in the coronal hole area data and predict its size for different sun regions over 7 days. By analyzing time series data on the coronal hole area, this study aims to identify patterns and trends in coronal hole behavior and understand how they may impact space weather events. This research represents an important step towards improving our ability to predict and prepare for space weather events that can affect Earth and technological systems.