Deforestation analysis using Random Forest and interactive supervised classification approach

The increasing demand for forest resources leads to overexploitation to a greater extent, raising the alarm for the environmental consequences. The excessive use of forest resources results in deforestation, which needs to be addressed for maintaining the natural ecosystem balance. This research focuses on the deforestation analysis of a territory in the Tartar district in Azerbaijan. The changes in the forest area have been analyzed using the high-resolution Azersky satellite datasets for the three alternative years, 2017, 2019, and 2021. Two classification approaches, namely Random Forest (RF) and NDVI- based interactive supervised classification, were implemented for this purpose. The statistical analysis of the results indicates the gradual decrease in the forest area from the year 2017 to the year 2021, which has been evaluated by visual interpretation through the change maps of the forest area. From RF classification results, it has been observed that there has been an overall decrease of 9.5% from 2017 to 2021. Also, the NDVI-based interactive supervised classification approach indicates an overall deforestation rate of 4.79% from 2017 to 2021. This work shows that the forest area in the study region has considerably reduced over the years, and there is a need to closely monitor deforestation in the considered study area.