A Study On Surges: I. Automatic Detection Of Dynamic Hα Dark Features From High-Cadence Full-Disk Observations

Abstract

We present a new method for the automatic identification and classification of dynamic Hα dark features found in time series of full-disk solar images at three Hα wavelengths (center, and ± 0.8 A). The simultaneous Hα observations are obtained by the multi-channel Flare Monitoring Telescope (FMT) at Hida Observatory. The program was developed in order to replace the present visual detection and classification of the phenomena. Usually, an obvious dark feature found in the Hα −0.8 A observations probably corresponds to some phenomenon such as a surge or chromospheric network enhancement, or filament activity. Thus, one of our aims in this program is to distinguish each phenomenon by its own properties and key parameters. We optimized the threshold values of the key parameters such as the area and darkness of the transiently darkening features in Hα −0.8 A so that the computer can reasonably identify surges and filament activations. In comparison, for a 7-day observation period, the number of dark events detected by the program contains 89% of the events recognized visually. However, 10 times more events are detected automatically. The missing events are mainly caused by the deletion of data with poor visibility. It is found that the dark events can be identified with more precise starting and ending times by a machine than by a human. Some statistical studies of surges or other activities can be carried out based on the computer-produced database. With some modifications the program can be applied to monitor real-time dynamic features on disk, including flare ribbons.