Method and Algorithms of Forecasting the Seasonal Characteristics of Anthropogenic Impact Areas Using Long-Term Remote Sensing Data

摘要

A method for predicting the characteristics of areas subject to the anthropogenic impact according to the Earth's remote sensing data is proposed. The developed method is based on the identification of patterns using long-term periodic observations. These patterns are applied to the seasonal observations of the current year. The method is implemented in a set of algorithms and predictive models. An example of the use of the method of the agricultural yield forecasting is given. The training and validation of models for the prediction of crop yields based on the long-term spatial data on the state of vegetation are described. Different regions of the Russian Federation including the Arctic regions are considered.