APPLICATION OF DEEP LEARNING IN SENSING URBAN VEGETATION AND ASSESSING ITS ECOSYSTEM SERVICES 5. METHODS, PRACTICAL IMPLEMENTATIONS, AND RISKS (A REVIEW OF INTERNATIONAL LITERATURE)
V. A. Usoltsev1, V. P. Chasovskikh2
1Ural State Forest Engineering University, Yekaterinburg, Russian Federation
2Ural State University of Economics, Yekaterinburg, Federation
Keywords: deep machine learning, neural networks, urban pollution, carbon deposition, oxygen production, monetization of ecosystem services, comfort of the urban environment
Abstract
Currently, more than 55 % of the world’s population lives in cities, and this figure is projected to increase by 70 % to 2050. One of the reasons for the mass migration of rural residents to cities is their desire to improve the quality of life. However, the dense construction of high-rise buildings has led to a monstrous concentration of people, which, combined with the growth of urban transport, has reduced the quality of the living environment to a level that poses a threat to human health and life. Today, urban plantings are considered as a universal tool for solving urban problems in terms of achieving an optimal balance between the services provided and the existing risks. Urban vegetation, especially woody one, has numerous properties that improve the quality of the environment and preserve human health in urban environments. The review shows that artificial intelligence methods open up great prospects for solving complex problems of monitoring air quality in the urban environment, and uncertainty about the oxygen-producing role of urban areas is revealed. A technique for automatic mapping of urban trees is presented, and a spatio-temporal neural network training system for assessing the carbon deposition potential of urban forests is described. The possibilities of using machine learning in the assessment and monetization of ecosystem services in urban trees are shown. Based on the results of deep learning, an assessment of the comfort level of the urban environment was carried out in connection with the density of buildings, the degree of greening, and taking into account the results of a survey of residents.
|
