Abstract:

The boundary of urban built-up areas provides foundational information for urban studies and meets the requirements for urban pattern and urban spatial structure research. However, commonly used methods for identifying the boundary of urban built-up areas such as using remote sensing data of night-light and land use, cadastral data, and building coverage data, are limited in accuracy. To remedy this, this paper proposes to use POI (Point of Interest) data obtained from web maps, assuming that it well reflects the agglomeration of urban activities at higher precision. Based on the underlying connection between POI and the spatial distribution of urban activities, a new method called 'Densi-Graph' is proposed to identify the actual boundary of urban built-up areas with the contour map of the kernel density of POI , where the threshold value for the contour lines to make significant change from densely to loosely placed is picked, giving the boundary of urban and rural areas. Different contour structures for mono-centric, poly-centric and linear cities are discussed, whereby the Densi-Graph method using POI data is validated. The method is also used to study the boundaries of urban built-up areas in China's prefecture level cities. The relationships between the Density-Graph thresholds and the population and location of different cities are discussed. This study advances previous studies in presenting more reliable and objective data on the boundary of urban built-up areas.

Key words: boundary, urban built-up areas, POI, Densi-Graph, kernel density analysis, urban-rural spatial structure