Forecasting coastal systems based on satellite images

Abstract

In the Southern region of Russia, biotic, biological and anthropogenic factors are constantly in effect. To simulate various options for the development of biological and geophysical processes in marine and coastal systems, there is a need to develop and create non-stationary spatially heterogeneous interconnected mathematical models. For practical application of the models, real input data (boundary and initial conditions) and information on the initial parameters are required. This information can be obtained using spacecraft. This paper presents the developed software and algorithmic tools for recognizing space images, based on a combination of methods - local binary patterns (LBP) and neural network technologies. Initial data based on space images are entered into the computer model, which provide high accuracy in determining the state of coastal systems. This model can be used to predict possible changes in coastal ecosystems and develop strategies for their protection. The obtained research results open up significant prospects for preventing and reducing the negative consequences of adverse natural phenomena, including intensive “blooming” of water, reducing the calculation time by 20-30%, which provides specialists with a more rapid response to environmental changes. Thus, the research results open up new opportunities for improving the quality and effectiveness of environmental forecasts.