2021-09-09, 12:25–12:45, HUGOTech
The rapidly increasing amount of publically available remotely sensed data in recent decades has revolutionized large-scale research and context-informed decision making. However, these data are generally not freely available as homogenized products ready for analysis at continental (or larger) scales. This is widely observed with datasets generated by EO satellites, particularly those with optical sensors and those capable of high-resolution imaging, where the process of mosaicking imagery to produce a homogenous, cloudless dataset across a particular area of interest often grows increasingly cumbersome at larger scales.
This work describes a method of producing smooth and cloudless wide-area seasonal datasets from satellite imagery through overlapping pixel averaging weighted by distance from the suborbital track, alongside traditional mosaicking methods performed on images generated within the same orbital period. The method is shown to scale well to large areas through efficient use of cloud computing. Additionally, a dataset (published with an open license) is presented as a result of the described method, consisting of seasonal pan-european Sentinel-2 imagery mosaics (in the optical and NIR bands), produced in 30 meter resolution and spanning multiple years.
Luka Antonić, MultiOne Ltd.
Josip Križan has 20 years of professional experience in development and application of computational methods in the environmental field. His broad expertise includes advanced GIS, large-scale processing of remotely sensed data, mathematical modeling for environmental applications, database architecture, and general-purpose programming in a wide array of languages. He has participated in numerous National and European environmental/IT projects (e.g. Croatian Environmental Information System, Croatian Soil Information System, National forest inventorization project) and R&D projects (e.g. EUREKA E!3266, E!5460, PHIME, REDFAITH, CroClimGoGreen, Geo-harmonizer). Since 2009., up to the recent founding of his own company, he was engaged as Head of Mathematical modeling group in Gekom - Geophysical and ecological modeling Ltd., participating in numerous environmental modeling projects. His research activities and interests are focused on the application of Machine Learning (ML) methods on environmental problems. He is a co-author of 9 articles published in CC journals (Ecological modeling and Atmospheric Environment), most of them related to the application of ML on environmental problems, such as habitat mapping, hydrological regime problems, renewable energy resources mapping, climatological variables over a large region with complex terrain, air quality, among others.