This research addresses the problem of early detection of smoke and open fire on the observed territory by unmanned aerial vehicles. We solve the tasks of improving the quality of incoming video data by removing motion blur and stabilizing the video stream; detecting the horizon line in the frame; and identifying fires using semantic segmentation with Euclidean–Mahalanobis distance and the modified convolutional neural network YOLO. The proposed horizon line detection algorithm allows for cutting off unnecessary information such as cloud-covered areas in the frame by calculating local contrast, which is equivalent to the pixel informativeness indicator of the image. Proposed preprocessing methods give a delay of no more than 0.03 s due to the use of a pipeline method for data processing. Experimental results show that the horizon clipping algorithm improves fire and smoke detection accuracy by approximately 11%. The best results with the neural network were achieved with YOLO 5m, which yielded an F1 score of 76.75% combined with a processing speed of 45 frames per second. The obtained results differ from existing analogs by utilizing a comprehensive approach to early fire detection, which includes image enhancement and alternative real-time video processing methods.

Рассмотрены методы когнитивно/графического представления телеметрической информации. Выполнен анализ применимости существующих методов визуализации многомерных данных для мониторинга динамических систем реального времени со сложной иерархической структурой. В заключительной части работы представлена таблица, обобщающая результаты анализа исследуемых методов.

В настоящей работе проведено исследование предметной области, выполнен обзор существующих разработок в области построения систем мониторинга, контроля и диагностики подсистем космических аппаратов, в том числе, с использованием нейросетевого подхода. Теоретически исследованы пути реализации математического и алгоритмического обеспечения системы контроля и диагностики подсистем космического аппарата. Разработаны методические подходы, способы и методы решения технических задач по построению нейросетевой системы контроля и диагностики подсистем космического аппарата. Применение технологий искусственных нейронных сетей позволяет обнаруживать, классифицировать и прогнозировать ошибки, осуществлять многоуровневую диагностику подсистем космического аппарата и прогнозировать их дальнейшее поведение, тем самым увеличивая эффективность, скорость принятия решений и надежность работы узлов космического аппарата. Представлен метод графического представления временных последовательностей, позволяющий визуально классифицировать радиотехнический сигнал и обнаружить шум в этом сигнале. Предлагается формировать и ранжировать набор значимых признаков путем применения алгоритмов «Add» и «Del».

An analysis of modern work in the field of building fast-acting neurons and neural networks was carried out. The algorithm for setting up a multilayer neur

The method of cognitive graphical information presentation is developed, allowing to classify radiotechnical signals and estimate the degree of noise. The method is based on the construction of a features set ordered by information significance. In turn, informativeness is determined by the formal contribution of the feature to the quality of signal type recognition. The construction of cognitive images is carried out in several stages: 1) determination of numerical characteristics of typical signals, 2) ranking and selection of the most informative characteristics, 3) construction of cognitive graphic images visualizing a multidimensional vector of signal features, 4) operator's interpretation of cognitive images. The method of integral contour representation polar scan is used to construct cognitivegraphic images of signals. A total of forty informative parameters of the signal (features) are calculated. To increase the polar scan selectivity, the features are ranked in informativeness descending order by the Add and Del methods. The operation of signals subtraction defined over their informative parameters is introduced. With the goal to improve the visual recognition quality, monochrome halftones have been added to cognitive images. To improve perception of the contour representation of difference images color components have been introduced. The sensitivity of the cognitive images to substantial noise signals is expressed through changes in polar scan forms, tones and colorful presentation. The comparison of signals recognition quality by using metrics and polar scan visual recognition is provided. The recommendations are given to the decision-making operator on the type and noise degree of the radio signal in the final part.

The methods of cognitive graphical information presentation are considered to support decision-making in earth stations of command measurement systems. The work is performed on transforming...