Search academic texts

Information × Registration Number 0219U100534, 0116U001584 , R & D reports Title Development of methodology for assessment and mapping of biophysical parameters of agricultural crops using high-resolution satellite data popup.stage_title Head Tkachenko Oleksij M., Кандидат технічних наук Registration Date 19-02-2019 Organization National University of Life and Environmental Sciences of Ukraine popup.description2  The object of study — maps of biophysical parameters, optical satellite images. The main goal of the work is to develop a methodology for assessing and building geospatial regional models and mapping biophysical parameters of crops based on machine learning methods, analyzing time series data and the inductive approach using high resolution satellite data and ground-based observations based on JECAM-Ukraine test site. Research methods — correlation analysis and methods for identifying patterns in data (data mining, data science) machine learning methods; regression modeling of dependencies between satellite products and ground-based measurements. A regression analysis has been carried out of the dependence of biophysical products on high-resolution satellite data for selected land cover classes, respectively, for European projects LUCAS and CORINE. Regression models and maps of biophysical parameters were constructed using ground-based and satellite measurements. Dependencies between predictors derived from satellite images and ground data are estimated. Using the ensemble neural network model, a land cover classification map was obtained for the Kiev region for the growing season of 2018. The integrated vegetation characteristics such as LAI and FAPAR were simulated using ground data and satellite data (Landsat-8, Sentinel-2, etc.). The relationships between satellite data and biophysical parameters are investigated. The assessment of the quality of the constructed models was carried out. A correlation analysis of the obtained results and satellite data for the corresponding periods of time (NDVI index) was carried out. Methods for land cover classification based on satellite images of high spatial resolution, as well as a methodology for estimating biophysical parameters using ground-based and satellite measurements, are proposed. Product Description popup.authors popup.nrat_date 2020-04-02 Close