Deforestation Mapping Using MODIS Tree Cover Mask and Sentinel-1 Images

Forests are an essential resource that needs to be conserved. Excessive cutting of trees for urbanization and growth has led to rapid deforestation in parts of Haryana and its neighboring areas. In this study, tree cover mapping is done over a period of five years (2015–2019) using Sentinel-1 ground range detected band-C images. To distinguish between the land cover classes, a rich set of features play an important role. Based on the second-order statistics, gray level co-occurrence (GLCM) features are extracted from the image to study the uniformity between the pixels. A binary classification of the study area into tree and non-tree area is carried out by supervised random forest algorithm. According to the analysis, the net rate of reduction of the tree cover in parts Haryana and its neighboring areas, i.e., parts of New Delhi, is calculated as 3.1% in successive years.