Estimation of Annual Deforestation Using Random Forest

Forests play an important part in sustaining life on the planet. Forests not only protect against natural disasters but also maintain nutrient-rich topsoil. Due to population growth and people's lack of discipline, the forests are being destroyed gradually. Effects of deforestation are climate change, loss of habitat, increasing greenhouse gases, soil erosion and flooding. The seventh-most deforested nation in the world is Myanmar, according to the FRA (Forest Resource Assessment) for 2020. In this paper, the prediction of deforestation is estimated annually by using remote sensing and Geographic Information Systems (GIS) in Myanmar. Deforestation is one of the main reasons for causing natural disasters. There is a need to manage natural disasters to protect the environment. To reduce climate change, it is need to know the state of deforestation. The proposed system uses popular machine learning algorithms such the Random Forest classification and regression techniques to determine deforestation and forest degradation. The system performs semantic segmentation on satellite images to determine the percentage of forested areas and the rate of change in deforestation between time intervals annually. Conserving the environment by assessing deforestation will help a lot to reduce climate change in Myanmar.