Comparison of machine learning algorithms for land cover classification from Sentinel-2 satellite images on the Google Earth Engine platform

Abstract

The Google Earth Engine cloud computing platform has proven highly effective in land cover classification. In this study, we utilized the Classification and Regression Tree (CART) and Random Forest (RF) algorithms to classify land cover in Sentinel-2 satellite images. The results in the study area showed significant variations between the two algorithms. Specifically, the CART algorithm achieved an overall accuracy (OA) of 0.92 and a Kappa coefficient of 0.85, while the RF algorithm had an OA of 0.89 and a Kappa coefficient of 0.86.