Solar flare detection using neural networks

R.A. FERNANDEZ BORDA, P.D. MININNI, C.H. MANDRINI, D.O. GOMEZ, O.H. BAUER, Y M.G. ROVIRA

Santiago, Chile

Workshop; Magnetic fields across the Hertzsprung-Russell diagram; 2001

ESO

We present a new method to automatically detect flare events in the optical range. The method uses neural networks for pattern recognition and is conceived to be applied to full disk H-alpha images. Images are pre-analyzed in real time, allowing the design of automatic patrol processes without human assistance, which can detect and record flare events with the highest time cadence available. In this work we use a two-layer network with a hidden layer of TLUs and an output layer of ADALINEs. The network was trained using a back-propagation algorithm. We have applied the method to full disk raw images obtained by HASTA (H-alpha Solar Telescope for Argentina), which is located at OAFA (Observatorio Astronomico Felix Aguilar), El Leoncito, San Juan, Argentina. This method is found to be useful for the detection of flares in the complete optical classification, being portable to any H-alpha instrument and providing unique criteria for flare detection independent of the observer.