Classification of urban land coverage using satellite data and random forest classifier

Satellite data provides great potential and opportunities for environmental applications. In addition, open-access data is an independent channel for obtaining environmental information. Numerous non-governmental environmental organizations in the world use satellite data for their own activities. In this study, the usefulness of the classification of geoinformation systems of remote sensing for the cartography of urban lands was investigated. The main purpose of the work is to study various vegetation areas in urban settlements and industrial zones, for example, in the Kryvyi Rih city. After completing the production of maps, a visual and field check of the classification accuracy of the buffer zones is performed. Some problems are not excluded due to small smoothing options. Numerous buffer zones between earth classes cause numerous small mosaics on the final map, but these mosaics provide additional information about the differences in the surrounding area.