Aachen 2019 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

T: Fachverband Teilchenphysik

T 60: Astroteilchenphysik: Methoden III

T 60.7: Talk

Wednesday, March 27, 2019, 17:30–17:45, S12

Identification of Cherenkov Signal in Shower Images Recorded by CTA — •Johan Wulff^1,2, Jonas Hackfeld^1,2, Julia Tjus^1,2, and Lenka Tomankova^1,2,3 for the CTA collaboration — ¹Ruhr-Universität Bochum, Theoretische Physik IV — ²Ruhr Astroparticle and Plasma Physics (RAPP) Center — ³Erlangen Centre for Astroparticle Physics (ECAP), Friedrich-Alexander-Universität Erlangen-Nürnberg

The Cherenkov Telescope Array (CTA) is the next-generation ground-based gamma-ray observatory surpassing current instruments by roughly an order of magnitude in sensitivity. This unprecedented sensitivity bears new challenges for the on-site analysis and brings about the necessity to reduce the recorded data stream by about two orders of magnitude to allow for efficient handling and off-site transfer. This process will take place on-the-fly and must not significantly deteriorate physics performance.

We present a data reduction approach based on identifying camera pixels containing Cherenkov signal and suppressing those containing only noise. The procedure is applied to individual air shower images, both gamma- and hadron-induced. We employ deep learning (DL) methods, in particular convolutional neural networks (CNNs), which show great pattern recognition in image analysis tasks. Following the introduction of the method itself, we discuss its application to and performance in data reduction for CTA, including a comparison with non-DL methods.