Estimating urban spatial structure based on remote sensing data

Sci Rep. 2023 May 31;13(1):8804. doi: 10.1038/s41598-023-36082-8.

Abstract

Understanding the spatial structure of a city is essential for formulating a spatial strategy for that city. In this study, we propose a method for analyzing the functional spatial structure of cities based on satellite remote sensing data. In this method, we first assume that urban functions consist of residential and central functions, and that these functions are measured by trip attraction by purpose. Next, we develop a model to explain trip attraction using remote sensing data, and estimate trip attraction on a grid basis. Using the estimated trip attraction, we created a contour tree to identify the spatial extent of the city and the hierarchical structure of the central functions of the city. As a result of applying this method to the Tokyo metropolitan area, we found that (1) our method reproduced 84% of urban areas and 94% of non-urban areas defined by the government, (2) our method extracted 848 urban centers, and their size distribution followed a Pareto distribution, and (3) the top-ranking urban centers were consistent with the districts defined in the master plans for the metropolitan area. Based on the results, we discussed the applicability of our method to urban structure analysis.