Программное обеспечение для автоматизированного распознавания и оцифровки архивных данных оптических наблюдений полярных сияний

One of the main tools for recording auroras is the optical observation of the sky in automatic mode using all-sky cameras. The results of observations are recorded in special mnemonic tables, ascaplots. Ascaplots provide daily information on the presence or absence of cloud cover and auroras in various parts of the sky and are traditionally used to study the daily distribution of auroras in a given spatial region, as well as to calculate the probability of their observation in other regions in accordance with the level of geomagnetic activity. At the same time, the processing of ascaplots is currently carried out manually, which is associated with significant time costs and a high proportion of errors due to the human factor. To increase the efficiency of ascaplot processing, we propose an approach that automates the recognition and digitization of data from optical observations of auroras. A formalization of the ascaplot structure is proposed, which is used to process the ascaplot image, extract the corresponding observation results, and form the resulting data set. The approach involves the use of machine vision algorithms and the use of a specialized mask - a debug image for digitization, which is a color image in which the general position of the ascaplot cells is specified. The proposed approach and the corresponding algorithms are implemented in the form of software that provides recognition and digitization of archival data from optical observations of auroras. The solution is a single-user desktop software that allows the user to convert ascaplot images into tables in batch mode, available for further processing and analysis. The results of the computational experiments have shown that the use of the proposed software will make it possible to avoid errors in the digitization of ascaplots, on the one hand, and significantly increase the speed of the corresponding computational operations, on the other. Taken together, this will improve the efficiency of processing ascaplots and conducting research in the relevant area.