Rapid Detection of Land Cover Changes Using Crowdsourced Geographic Information: A Case Study of Beijing, China

摘要: Land cover change (LCC) detection is a significant component of sustainability research including ecological economics and climate change. Due to the rapid variability natural environment, effective LCC required capture sufficient change-related information. Although such information has been available through remotely sensed images, complicated image processing classification make it time consuming labour intensive. In contrast, freely crowdsourced geographic (CGI) contains easily interpreted textual information, thus potential be applied for capturing Therefore, this paper presents evaluates method using CGI detection. As case study, Beijing chosen as study area, monitor one kind which generated from commercial Internet maps, points interest (POIs) with detailed are utilised detect land in 2016. Those POIs first classified into nomenclature based on their Then, kernel density approach proposed effectively generate regions Finally, GlobeLand30 2010 baseline map, detected post-classification period 2010–2016 Beijing. The result shows that an accuracy 89.20% achieved by POIs, indicating reliable Additionally, comparison between images CGI, revealing advantages terms efficiency. However, due uneven distribution, still areas few POIs.