Automatic Solar Flare Detective with the Solar Disk Imager onboard the ASO-S Mission

We present an automated solar flare detective, a software tool to automatically process solar observation images, detect and track solar flares, and finally compile an event catalog. It is capable of identifying and tracking flares that happen simultaneously or temporally close together. The method to identify a flare is based on the local intensity changes in macropixels. The basic characteristics such as the time and location information of a flare are determined with a triple-threshold scheme, with the first threshold (global threshold) to determine the occurrence (location) of the flare, and the second/third threshold (local threshold) to determine the start/end time of the flare. We have applied this tool to one month of continuous solar ultraviolet (UV) images obtained by the Solar Disk Imager (SDI) onboard the Advanced Space-based Solar Observatory (ASO-S), which show solar activities such as flare, filament/prominence, and solar jets. Our automated tool efficiently detected a total number of 226 solar events. After a visual inspection, we found that only one event was misidentified (not related to solar activities). We compared the detected events with the GOES X-ray flare list and found that our tool can detect 73\% of GOES M-class and above flares (37 out of 51), from which we conclude that the intensity increase in SDI UV images can be considered as a good indicator of a solar flare.