Search academic texts

Information × Registration Number 0226U000028, (0123U100935) , R & D reports Title Development of the theory of restoration of radio images of space sources and neural network methods of semantic segmentation of Earth remote sensing data. popup.stage_title Розвиток теорії відновлення радіозображень космічних джерел та нейромережевих методів семантичної сегментації даних дистанційного зондування Землі. Head Rusyn Bohdan P., Доктор технічних наук Registration Date 01-01-2026 Organization Physico-Mechanical Institute named after GV Karpenko of the National Academy of Sciences of Ukraine popup.description1 To develop the theoretical foundations of the restoration of radio images based on the fine structure of the response of the radio interferometric system, to investigate the set of dynamic helio-geophysical manifestations of solar activity, to develop new methods of semantic segmentation of objects on digital multispectral images using neural network technologies and criteria for evaluating the informativeness of training samples for the automation of machine learning, develop approaches to the integration of regional specialized geoinformation systems and remote sensing data of the Earth. popup.description2 The theoretical foundations for radio image restoration of space sources based on a tomographic approach for conditions specific to decameter waves have been developed. The influence of solar high-energy particles on atmospheric infrasound and atmospheric electric field during a proton event caused by interacting coronal mass ejections has been studied. Semantic relationships between objects presented in remote sensing images have been studied and new methods of semantic image segmentation have been developed for their detection and recognition in dual-purpose tasks. Criteria for assessing the informativeness of the training sample have been developed for automating machine learning and deep neural networks. Based on the results of contact and remote monitoring: an integration model has been created: regional GIS of the protected area + satellite data (Sentinel-2, Landsat 8/9) + GoogleEarthEngine); various levels of integration have been implemented: at the data level (NDWI, NDTI, NDCI, NDSI); at the level of analytical processes (time series, correlations); climate scenarios (NASA NEX-GDDP, rcp45/rcp85); comprehensive monitoring provided: 26 lakes (water level, water quality); forest ecosystems (condition, degradation risks, carbon balance); climate risks (temperature, precipitation, scenarios until 2050). Product Description popup.authors Rostyslav J. Kosarevych Oleksiy A. Lutsyk Bohdan P. Rusyn Oleh L. Ivantyshyn Andrij B. Lozynsʹkyy Anatoliy M. Zavediya Mykola M. Korus Galyna M. Statkevych Natalya А. Pitsʹ Borys S. Kharchenko Dariya V. Ivchenko Olha V. Alokhina popup.nrat_date 2026-01-01 Close