Peatlands store and emit large amounts of greenhouse gases. With the climate changing due to global warming, measuring these emissions helps to get a better understanding of the role of peatlands in the global carbon cycle. Measurements at a bog site of the Siikaneva peatland show that the emissions vary along the different microtopographies shaped by their vegetation and ground water level. To upscale these measurements, a supervised classification of the study area was implemented in this study by testing a method that uses high-resolution multispectral aerial imagery, captured by a UAV (Uncrewed Aerial Vehicle), and a Random Forest classifier. A cohesive orthomosaic of the study area was produced, training data were generated to adjust the Random Forest model, and the study area was classified. The results show that the applied methods were successful in generating a multispectral orthomosaic as well as a classified raster of the study area. A mean classification accuracy of 75.7 % was achieved, which can be considered as a good result. Misclassification rates of neighboring microtopographies with similar vegetation could be mitigated by utilizing a LiDAR (Light Detection and Ranging) sensor in further studies.