Comparison of Different Classification Approaches for Land Cover Classification using Multispectral and Fusion Satellite Data: A Case Study in Ören Forest Planning Unit

In this study, the success of different satellite images and classification approaches in land cover (LC) were compared. A total six images, including two passive (Landsat 8 OLI (L8) Sentinel-2 (S2)) four fused from active (Sentinel-1(S1)-VH VV polarization) (L8-S1-VH, L8-S1-VV, S2-S1-VH S2-S1-VV) used study. For purpose, L8, S2, L8-S1-VH, S2-S1-VV classified according to three ((Maximum Likelihood Classification (MLC), Support Vector Machine (SVM) Artificial Neural Networks (ANN)) image using forest types map as gorund data. The results obtained methods evaluated based on overall accuracies (OA) kappa coefficients (KC). When successes are evaluated, it was observed that KC ranged 0.66 0.95 OA 76.82% 96.67. indicated highest displayed by MLC (ranged 85.33% 96.67%), closely followed SVM 80.11% 91.93%), finally ANN 89.92%). addition, a comparison performance utilized algorithms performed. S1-VH; S1-VV and, S2 L8 with an algorithm produce most accurate LC map, indicating 92.00%, 94.00%, 96.67%, 93.33% 0.90, 0.93, 0.95, 0.92 for respectively. Thus, can be concluded improve classification.