A COMPARATIVE ANALYSIS OF PIXEL-BASED AND OBJECT-BASED APPROACHES FOR FOREST ABOVE-GROUND BIOMASS ESTIMATION USING RANDOM FOREST MODEL

Abstract. Providing an accurate above-ground biomass (AGB) map is of paramount importance for carbon stock and climate change monitoring. The main objective this study to compare the performance pixel-based object-based approaches AGB estimation temperate forests in north-eastern New York State. Second, capabilities optical, SAR, optical + SAR data were investigated. To achieve goals, random forest (RF) regression algorithm was used model predict values. Optical (i.e. Landsat 5TM, 8 OLI, Sentinel-2), synthetic aperture radar (SAR) (Sentinel-1 global phased array type L-band (PALSAR/PALSAR-2)), their integration have been estimate AGB. It worth mentioning that airborne light detection ranging (LiDAR) raster has as a reference training/testing purposes. results demonstrate OBIA approach enhanced RMSE about 5.32 Mg/ha, 8.9 5.29 Mg/ha data, respectively. Moreover, with 42.63 R2 0.72 37.31 0.77 provided best results.