07. PHÂN LOẠI ĐỐI TƯỢNG CHIẾT TÁCH LỚP PHỦ BỀ MẶT TẠI KHU VỰC CÔNG VIÊN ĐỊA CHẤT TOÀN CẦU NON NƯỚC CAO BẰNG DỰA TRÊN NỀN TẢNG ĐIỆN TOÁN ĐÁM MÂY

Object - based land cover classification on the cloud computing platform at Non Nuoc Cao Bang global geopark

Abstract

Land cover is one of the most meaningful input factors for geopark management and monitoring. Previously, extracting of land cover data from remote sensing images has used commercial softwares. Due to the limitations of computer hardware and algorithms, commercial softwares increase the time and cost of mapping. The arrival of cloud computing platform Google Earth Engine (GEE) in 2010 has brought a breakthrough for analyzing and processing remote sensing images. Therefore, in this article, cloud computing technology is studied to build land covers in 2019 for Non Nuoc Cao Bang global geopark area. The classification results comprise 6 types of land covers, including: paddy, rural area, artificial forest, natural forest, water and other cultivated land. The classification accuracy is relatively high, overall accuracy is 83.2%, Kappa coefficient is 0.78. This classification results contribute significantly to the management and monitoring tasks in the geopark area.