Digital image processing for urban sprawl classification in google earth

Abstract

Maps of land use and land cover are necessary for studying how the earth's surface changes over time and how human activities affect their surroundings the expanding pool of available resources of remote sensing, particularly the wellarchived Sentinel-2B images with resolution of 10 metres were used to conduct the research. For land use and land cover maps, supervised categorization is used. However, ground truth is required to attain high classification accuracy. There is a need for high-quality samples in big quantities. It takes time and effort to collect ground truth samples. When it comes to ground truth, it might be costly and sometimes impossible to obtain. This work presented a method for using Google Earth Engine products as ground truth instead of manually labelling ground truth samples and proceeding with classification with region of interest to produce land use and land cover maps for 2021 in this study area, Amravati, on the Google Earth Engine platform in this paper. The accuracy test is carried out on randomly generated samples from various places, with an overall accuracy of roughly 83.4 percent.