An integrative method for mapping urban land use change using "geo-sensor" data

Due to lack of high-resolution data both in space and time, mapping land use change in built-up areas remain challenging. In this work, we developed an approach that integrates multiple "geo-sensor" data sources, including remote sensing and social media, for mapping urban land use changes. The approach starts by mapping built-up expansion annually in small patches, using dense time periods of remotely sensed imagery. We refine these patches and identify the major categories of land use such as residential, commercial, manufacturing, recreational, and office. We further select one major category---residential---and use social media data and house trading records to classify it into three sub-categories: gated communities, ordinary communities, and urban slums. We demonstrate our approach using the multi-sourced data of Kunming, a medium-sized city in China. The results showed that our approach provides an efficient and integrative way for mapping large-scale land use change in urban regions with significant details that are not feasible only with limited remote sensing data.