Алгоритм и оценка изменения площади лесного покрова Хачмазского района Азербайджана средствами космического мониторинга

Abstract

The article presents the results of long-term observations of the forest cover state in the Khachmaz region of Azerbaijan. Both the results of Earth’s remote sensing and the data of earlier observations carried out by field monitoring were used in the study. The earlier data was stored in the form of thematic maps on paper. The results of remote sensing contained multispectral satellite images obtained over several years. Image processing was performed using the QGIS program and the Semi-Automatic Classification Plugin. The processing included radiometric and atmospheric correction, as well as classification of the study area by the type of land surface coverage. Classification was performed on the basis of the spectral curves analysis in various sections of the region. The areas occupied by forests were determined based on the classification. The set of observations included data from the available archival materials – thematic maps of the Khachmaz region. A software in MATLAB was developed for processing the maps and calculating the area of forests. The program operation algorithm includes histogram analysis of the image as a whole and separately of the legend. The histogram of the legend determines the number of thematic layers of the map. They do not include layers containing unused colors, such as the colors of neighboring areas, sections of the sea, etc. Then color correction of image pixels is performed, namely, quantization of colors according to the number of specific thematic layers. Later on, morphological processing of each layer is carried out. Areas containing less than the specified number of pixels are assigned the color of the surrounding layer. The number of pixels of each layer and all layers are calculated. Finally, the obtained ratios are used to calculate the area of each layer. The data of processing of archival materials together with the results of remote sensing are summarized in a joint table. According to the table, a diagram of changes in the forest area is constructed, as well as a model in the form of a polynomial showing this dynamics. The diagram analysis revealed a tendency to a decrease in the area of forests. The forest area has decreased by 21 % for 7 years. The developed algorithm splits thematic maps into separate layers in accordance with the colors of the legend.