Comparison of the dynamics of coronal holes identified by two detection methods in solar activity cycles 24-25

In this study, we analyze for the first time the long-term variations of coronal holes (CHs) identified by two different automatic detection schemes: Spatial Possibilistic Clustering Algorithm (SPoCA) and Convolutional Neural Network (CNN193). The source material was the observational data acquired by the Atmospheric Imaging Assembly instrument onboard the Solar Dynamics Observatory (AIA/SDO) in the Fe XII 19.3 nm line in the period from June 16, 2010, to May 13, 2021. An initial analysis comparing the long-term variations of the CH areas extracted by the SPoCA method and used by us in earlier works to study their evolution at different stages of the 24th and at the beginning of the 25th cycles showed fairly good agreement with the trend of the CH areas identified by the CNN193 algorithm for the same period. Both schemes reveal hemispheric asymmetry in the generation of CHs both in time and in amplitude.