The Cherenkov Telescope Array (CTA) represents the next generation of ground based observatories for very high energy gamma ray astronomy. The CTA will consist of two arrays at two different sites, one in the northern and one in the southern hemisphere. The current CTA design foresees, in the southern site, the installation of many tens of imaging atmospheric Cherenkov telescopes of three different classes, namely large, medium, and small, so defined in relation to their mirror area; the northern hemisphere array would consist of few tens of the two larger telescope types. The telescopes will be equipped with cameras composed either of photomultipliers or silicon photomultipliers, and with different trigger and read-out electronics. In such a scenario, several different methods will be used for the telescopes’ calibration. Nevertheless, the optical throughput of any CTA telescope, independently of its type, can be calibrated analyzing the characteristic image produced by local atmospheric highly energetic muons that induce the emission of Cherenkov light which is imaged as a ring onto the focal plane if their impact point is relatively close to the telescope optical axis. Large sized telescopes would be able to detect useful muon events under stereo coincidence and such stereo muon events will be directly addressed to the central CTA array data acquisition pipeline to be analyzed. For the medium and small sized telescopes, due to their smaller mirror area and large inter-telescope distance, the stereo coincidence rate will tend to zero; nevertheless, muon events will be detected by single telescopes that must therefore be able to identify them as possible useful calibration candidates, even if no stereo coincidence is available. This is the case for the ASTRI telescopes, proposed as pre-production units of the small size array of the CTA, which are able to detect muon events during regular data taking without requiring any dedicated trigger. We present two fast algorithms to efficiently use uncalibrated data to recognize useful muon events within the single ASTRI camera server while keeping the number of proton induced triggers as low as possible to avoid saturating the readout budget towards the central CTA data analysis pipeline.