Meteor detection from space with Mini-EUSO telescope

<p>Mini-EUSO is a very wide (44°x44°) field of view telescope installed on August 2019 inside the Zvezda Module of the ISS, looking nadir through a UV transparent window and taking data since October 2019. Its optical system consists of two Fresnel lenses, focusing the light onto an array of 36 multi-anode photomultiplier tubes. The focal surface counts a total of 2304 pixels, each one having a footprint of about 6.5 km on ground. The instrument triggers on two different timescales, respectively 2.5 μs (D1) and 320 μs (D2), and perform a continuous monitoring of the UV emission at a 40.96 ms timescale (D3). At time of writing, about one thousand meteors on D3 data have been classified as meteors using our current detection algorithm. We describe here a concept of an alternative algorithm to recognize meteors in the D3 continuous data-stream, which can be also implemented in the future for online triggering, and show some examples of detected meteors by our instrument. We also performed a search of possible coincident detections of Mini-EUSO meteors by ground meteor and fireball networks, such as PRISMA in Italy, to gain a stereoscopic vision of the event itself. In light of these initial results, we present here the capabilities of Mini-EUSO instrument in meteor science.</p>

First observations of Polar Mesospheric Echoes at both 31 MHz and 53.5 MHz over Svalbard (78.2°N 15.1°E)

During summer 2020, observations of the mesosphere using a 53.5 MHz radar on Svalbard, at 78.2°N 15.1°E, revealed the well-known Polar Mesospheric Summer Echoes (PMSE). At the same time, a co-located meteor detection radar, operating at 31 MHz detected corresponding echoes very distinct from those associated with meteor trails. Comparing as many days as possible during 2020, incontestable evidence arose to demonstrate that the meteor detection radar was capable of observing PMSE, although not in the optimised fashion of the 53.5 MHz system. We present examples of results from both systems, supplementing the earlier findings of Swarnalingam et al. (2009), and simultaneously show very first results from this particular geographical location.

AVOIDING MURPHY'S LAW ON DETECTING METEORS

It is known that when a meteoroid has been imaged by two or more stations, its atmospheric trajectory can be inferred. In addition to this, if the velocity of the meteor has been measured, then the magnitude, the photometric mass and its orbital elements can be computed. Hence, meteor detection networks have a large number of stations. Unfortunately, weak meteors are only imaged by the nearest station, since the brightness decreases with the square of the distance. On the other hand, Murphy’s law can act in the event of brilliant meteors and fireballs: “In a station it was cloudy. In another, the fireball was hidden under the horizon. A third was out of order due to an electrical power failure, and the other was under maintenance, etcetera.” Do not panic. In this work we present some methods to obtain information from a meteor seen from a single station, if it has been possible to associate it with a meteor shower. In this work, CCD images gathered by the robotized networks of the Sociedad Malague˜na de Astronom´ıa (Aznar 2016) and the BOOTES-1 and -2 observatories have been used (Castro-Tirado et al. 2008).

Deep Transfer Learning for Meteor Detection

In this paper, a pre-trained deep Convolutional Neural Network is applied to the problem of detecting meteors. Trained with limited data, the best model achieved an error rate of 0.04 and an F1 score of 0.94. Different approaches to perform transfer learning are tested, revealing that the choice of a proper pre-training dataset can provide better off-the-shelf features and lead to better results, and that the use of very deep representations for transfer learning does not worsen performance for Deep Residual Networks.