Geoinformation technology for land use dynamics monitoring based on the analysis of satellite images

• Datum: 2025
• Hauptverfasser: Okhariev, Viacheslav, Pidsadnii, Serhii
• Format: Artikel
• Sprache: Ukrainian
• Veröffentlicht: Kyiv National University of Construction and Architecture 2025
• Schlagworte: geographic information technology; land use; satellite image; error-adjusted data estimation; bootstrap analysis
• Online Zugang: https://es-journal.in.ua/article/view/343568
• Назва журналу: Environmental safety and natural resources

Changes in land cover have a significant impact on the global climate balance. Rational and balanced use of land resources is one of the key factors in reducing the negative impact of economic activity on climate change. Modern methods of remote sensing of the Earth allow for a comprehensive analysis of land use processes. Today, there are a number of products that provide land use classification based on open high-precision satellite images, in particular the Dynamic World (DW) dataset with a spatial resolution of 10 m. The purpose of this study is to assess the dynamics of land use in Ukraine, namely changes in the areas of vegetation classes and those used for growing crops, which may contribute to an increase in CO2 emissions into the atmosphere. This study analyzed Dynamic World land use classifications for the territory of Ukraine for 2017 and 2024. Despite the accuracy declared by the company at 73%, a separate reliability assessment was performed specifically for the territory of Ukraine. Based on a stratified sample, it was determined that the overall accuracy of the classification for 2024 is 51.15%, and the Kappa coefficient is 0.44, which is significantly lower than the officially published indicators. Taking into account the specifics of the territory, the nature of land use and the limited sample for 2017, an analysis of the dynamics of land use areas was conducted. Two groups of classes were considered: on the one hand, lands with different types of vegetation cover (trees, grasses, wetlands, shrubs), on the other hand, territories of active agricultural use with open soils (crops). An error-adjusted area estimation was carried out in the interaction of the crops class and classes with vegetation cover, and the results were corrected using a confidence interval and bootstrap analysis. This allowed us to take into account the impact of the reduced accuracy of DW classifications and the limitations of the stratified verification sample. The results showed a potential trend of the transition of classes with vegetation cover to the category of open soils, which may be associated with an increase in the intensity of agricultural land use. This, in turn, creates risks for the balance of emissions and assimilation of greenhouse gases.