O-LCMapping: a Google Earth Engine-based web toolkit for supporting online land cover classification

Land cover classification is essential for environmental monitoring, sustainable development goals assessment and other fields. Traditionally, it often takes much time and labor costs to copy and pre-process remote sensing images for performing land cover classification with desktop software, which must be installed and licensed. Google Earth Engine (GEE), as a cloud computing platform, provides a large number of remote sensing images, geoprocessing algorithms and massive computational capabilities via web interfaces. However, GEE requires users to possess programming skills of JavaScript or Python language, which can hinder the enthusiasm of many academics without programming skills. This paper developed a web-based and open-access toolkit O-LCMapping for supporting online land cover classification. This toolkit was implemented with JavaScript application programming interfaces of GEE and integrated ten imagery classification algorithms. The toolkit explicitly provides the entire pipeline of land cover classification in the form of user interfaces for end-users with basic knowledge of remote sensing but little programming skills. Concretely, it enables end-users to define study area, select remote sensing data, mark classification samples, set classification algorithms, evaluate accuracy and output results through user interfaces. Three different experimental cases indicate that the toolkit can be easily applied to different fields for various applications of land cover classification.