Deep learning methods and models for applied problems of satellite monitoring

Project within the NRFU competition “Leading and Young Scientists Research Support”

Project duration: 2020-2022

Supervisor: prof. Nataliia Kussul

Project code: 2020.02/0292

Project abstract

For applied big data tasks solving purpose, in the last years, intelligent information processing methods and deep learning methods have been actively developed. However, modern deep learning methods are developed mainly for work with textual information and images. This why, deep learning approaches for work with satellite data have to be developed and improved, in cause of the amount of these data has been growing exponentially in recent years. Therefore, the project will solve problems relevant to the modern world science – the development of the new recurrent deep learning models, generative adversarial networks and the fusion methods for the multidimensional geospatial data, which will harmonize satellite data with different nature, spatial and temporal resolution. This will allow creation of the high-level analytical products that will be able to meet the needs of economic indicators monitoring and complex agro-monitoring tasks.

Leave a Reply

Your email address will not be published. Required fields are marked *