08. Research classification of land use - land cover from Sentinel - 2 satellite images based on the SVM algorithm

Abstract

The Sentinel - 2 satellite imagery is a free source with a high spatial and temporal resolution, this data has to be effective in classifying land use - land cover for monitoring, and land management. Based on the Google Earth Engine (GEE), this study uses the Support Vector Machine (SVM) algorithm applied to machine learning to classify the land use - land cover status in Bac Tu Liem district, Hanoi. The results of the study show that there are 4 basic types of land use - land cover, in which the water - bodies area is the smallest with 422.31 hectares and the largest area is the cover of Vegetation with 1876.97 hectares. The classification accuracy achieved good results with the Kappa coefficient reaching 0.82.