GOOGLE EARTH ENGINE FOR LANDSAT IMAGE PROCESSING AND ASSESSING LULC CLASSIFICATION IN SOUTHWESTERN CÔTE D’IVOIRE

High-accuracy land use and cover maps (LULC) are increasingly in demand for environmental management decision-making. Despite the limitation, Machine learning classifiers (MLC) fill gap any complex issue related to LULC data accuracy. Visualizing land-cover information is critical mitigating Côte d’Ivoire’s deforestation planning using Google Earth Engine (GEE) software. This paper estimates probability of RF classification South Western d’Ivoire. Landsat 8 Surface Reflectance Tiers 1 (L8OLI/TIRS) with a resolution 30 mn 2020 were used classify western southwestern Forest areas The Random (RF) classifier was calibrated 80% training 20% testing assess GEE accuracy performance. findings indicate that class accounts 39.48% entire study area, followed by Bareland class, Cultivated 21.28±0.90%, Water 1.94±0.27%, 0.96±0.60% Urban respectively. reliability test results show 99.85%±1.95 overall (OTA), 99.81±1.95% kappa (TK). validation (VOA) 94.02±1.90%, while 92.25±1.88% (VK) 92.45±1.88% Accuracy. different coefficients obtained from confusion matrix each has three good performances. due cultivated samples lower spatial smaller sample numbers, resulting PA this than other classes. All had producer (PA) user (UA) more 90% L8OLI/TIRS data. Using RF-based method integrated into provides an efficient high scores classifying area.