Search academic texts

Information × Registration Number 0220U000297, 0118U005269 , R & D reports Title Intelligent technologies for satellite monitoring of the environment based on in-depth training and cloud computing (InTeLLeCT) popup.stage_title Head Kussul Nataliya, Registration Date 28-01-2020 Organization Space Research Institute popup.description2 The object of study - satellite images of high spatial resolution and methods of their processing. The purpose of the work is to explore the development of the latest environmental monitoring technologies based on large amounts of satellite data, intelligent geo-information technologies, cloud computing and the Semantic Web. Research method - machine learning, geospatial analysis, object-oriented design, structural functional analysis, statistical methods, neural networks. The basic methods of machine learning are considered in the paper and the features and results of their application to the problems of classification of the earth cover according to satellite data of high resolution are analyzed. Ground data were collected for the territory of Kyiv region from 2018 to 2019. The obtained data are presented in digital form as polygons for different classes of the earth's surface. Available satellite images on the territory of Kyiv region for the growing seasons 2016-2019 are analyzed. Modern sources of optical range and radar data are analyzed. As a result of the analysis, it became clear that the Sentinel-1 and Sentinel-2 satellite data could be used to address the project's objectives. As a result of the analysis, it became clear that the Sentinel-1 and Sentinel-2 satellite data could be used to solve project objectives. Particular attention is paid to the deep architecture that today is the most powerful and accurate method to detect visual images. The main advantages of deep learning methods over the traditional approaches to classification problems, which have been used over the last decades and are based on expert knowledge to distinguish features from input, are identified. Ground cover products were created based on collected terrestrial and satellite data for the territory of Kyiv region. Product Description popup.authors Адаменко Т.І. Колотій А.В. Куссуль Н.М. Лавренюк М.С. Підгородецька Л.В. Тютюнник Л.І. Шелестов А.Ю. Шуміло Л.Л. Яйлимов Б.Я. popup.nrat_date 2020-04-02 Close