Адаптивно‑многоиндексно‑кластерный алгоритм комплексной оценки воздействия химических загрязнений на лесные массивы по спутниковым фотоизображениям

Abstract

An original adaptive-index-clustering algorithm is proposed: “Managed vegetation index”. An original adaptive-multi-index-cluster algorithm for comprehensive assessment of the impact of chemical pollution on forests using satellite photographs is proposed, which is distinguished by the use of an adaptive procedure for the formation of pixel clusters displaying a plurality of spectral channels of a photographic image of each type of vegetation state of a forest stand in the zones of chemical pollution of forest tracts, as well as using the procedure for calculating the weighted average values of complex vegetation indices for each zone of chemical pollution, which allows, based on the values of complex vegetation indices, to determine various biological, phytological and physico-chemical states of forest areas.It should be noted that in order to solve the complex problem of constructing complex indices linked to ecological zones, it is proposed to use the simple idea of increasing the quality of modeling and forecasting by expanding the amount of information. The proposed problem can be solved using a statistical analysis of data on the distribution of pixels whose belonging to ecological zones is known in advance. The development of the algorithm is based on the following prerequisites: (1) using a linear combination of individual classical vegetation indices of the state of forest areas, it is possible to create a new specialized complex vegetation index that makes it possible to identify ecological zones in forest areas according to the levels of impact on forests of chemical pollution of industrial enterprises; (2) the possibility of using specialized complex vegetation indices in the form of weighted average linear combinations of classical vegetation indices. Specialized complex vegetation indices of adaptive selection of weight coefficients are capable of displaying various biological, physicochemical and ecological characteristics of the state of forests based on clustering of satellite image pixels. The proposed algorithm makes it possible to calculate, as a result of clustering, more accurate estimates of the total areas of ecological zones of forest tracts, which can be used as a basis for assessing the degree of ecological degradation of forest tracts and environmental damage.