An Improved Method for Urban Built-Up Area Extraction Supported by Multi-Source Data

Urban built-up areas, where urbanization process takes place, represent well-developed areas in a city. The accurate and timely extraction of urban built-up areas has a fundamental role in the comprehension and management of urbanization dynamics. Urban built-up areas are not only a reflection of urban expansion but also the main space carrier of social activities. Recent research has attempted to integrate the social factor to improve the extraction accuracy. However, the existing extraction methods based on nighttime light data only focus on the integration of a single factor, such as points of interest or road networks, which leads to weak constraint and low accuracy. To address this issue, a new index-based methodology for urban built-up area extraction that fuses nighttime light data with multisource big data is proposed in this paper. The proposed index, while being conceptually simple and computationally inexpensive, can extract the built-up areas efficiently. First, a new index-based methodology, which integrates nighttime light data with points-of-interest, road networks, and the enhanced vegetation index, was constructed. Then, based on the proposed new index and the reference urban built-up data area, urban built-up area extraction was performed based on the dynamic threshold dichotomy method. Finally, the proposed method was validated based on actual data in a city. The experimental results indicate that the proposed index has high accuracy (recall, precision and F1 score) and applicability for urban built-up area boundary extraction. Moreover, this paper discussed different existing urban area extraction methods, and provides an insight into the appropriate approaches selection for further urban built-up area extraction in cities with different conditions.