"The article proposes a method for assessing the state of natural objects that is based on the complex use of heterogeneous models and hyperspectral Earth remote sensing data used to estimate the parameters of these models. The base model is an artificial neural network, for the training of which multiparametric models of radiation transfer, gradient search algorithms, as well as regression empirical models supplementing them, can be used and adaptively adjusted. The advantage of the method is the ability to determine the state of water bodies and vegetation under conditions of uncertainty with the possibility of making more precise estimates for the limited volume of ground measurements at reference points. Examples of approbation of the method in determining the state of coastal waters of the Black Sea with the use of hyperspectral imaging materials from “Resurs-P” satellite, as well as in assessing the state of vineyards, are shown."
Index of Natural Objects
