The method of forest change detection using Sentinel-2 optical satellite imagery and Sentinel-1 radar imagery: A case study in Dak Nong Province, Vietnam

Abstract

This study presents the findings of forest change detection in Dak Nong Province, Vietnam by combining the NDVI of Sentinel-2 satellite imagery and the backscatter (BKS) values of VH and VV polarizations from the first quarter of 2020 to the first quarter of 2021 and using the combination model of BKS and NDVI (CMB) to determine the NDVI and Backscatter Change Index (NBCI) between the two periods, which were processed and analyzed in Google Earth Engine. The verification results on 270 samples for all 3 indices show that the results of forest loss detection using NBCI index, NDVI index, and BKS index reach 93.3%, 83.3% and 73.3%, respectively; the results of identifying the areas with no forest change using NBCI index, NDVI index, and BKS index achieve 97.8%, 92.2% and 83.3%, respectively; the results of forest increase detection using NBCI index, NDVI index, and BKS index reach 90.0%, 85.6% and 68.9%, respectively. The NBCI index for Dak Nong Province indicated that during the study period, the the province lost 1,198.9 hectares forest area, whereas it gained 871.3 hectares of the same.