An accurate fringe extraction model of small- and medium-sized urban areas using multi-source data

Abstract

Urban fringes are of great significance to urban development as connecting hubs between urban and rural areas. However, there are many problems in urban fringes, including disorderly spatial layout, waste of social resources, and low quality of human settlements. Rapid and accurate identification of urban fringes has important practical significance for optimizing urban spatial layout, controlling urban unlimited expansion, and protecting land resources. Given the lack of suitable and high-quality fringe extraction models for small- and medium-sized urban areas, this study was based on Gaofen-2 (GF-2) imagery, Suomi National Polar-orbiting Partnership Visible Infrared Imager Radiometer Suite (NPP-VIIRS) imagery, point of interest (POI) data, and WorldPop data, taking the landscape disorder degree, POI kernel density, and night light intensity as urban feature factors and constructing a fringe extraction model of small- and medium-sized urban areas (FEM-SMU). Taking Hantai District in China as the study area, the results of the model were compared to the landscape disorder degree threshold method and POI kernel density breakpoint analysis method, while the generality of the model was further tested in Shangzhou and Hanbin Districts. The results show that the FEM-SMU model has evident improvements over the conventional methods in terms of accuracy, detail, and integrity, and has higher versatility, which can better meet the research needs of small- and medium-sized urban fringes.