.png){kind=icon}
.png){kind=icon}
Return
Enhancing the accuracy of urban population spatialization at 100-m grid scale through differentiating residential and non-residential buildings
.png){kind=icon}
PRE
AI
T
Q
Y
DOI：10.3389/feart.2026.1748599.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Population grid data have significant implications for socioeconomic development, urban planning, and environmental protection. However, current mainstream gridded population data products (e.g., WorldPop and LandScan) at the 100-m grid scale often fail to adequately capture the actual population distribution within urban physical areas. This inadequacy is mainly due to the lack of integration of three-dimensional (3D) building information and the failure to distinguish between residential and non-residential areas within cities, both of which make it difficult to reflect the differences in population agglomeration caused by variations in building volume and population types within the same grid. This paper distinguishes between residential and non-residential buildings and rasterizes the population onto residential buildings. First, the study utilized GF-2 remote sensing imagery and building vector data to identify residential structures within urban physical areas, followed by the calculation of building volumes using Digital Surface Model (DSM) data. Second, at the street-unit scale, a statistical relationship model was established between the residential building volume and the number of permanent residents. Finally, this relationship model was used to estimate and simulate the distribution of urban permanent residents at the 100-m grid resolution. The research findings reveal a strong correlation between the number of permanent residents and the residential building stock. The R2 values of the estimated models for the years 2010 and 2020 were 0.86 and 0.91, respectively. Compared with the 100-m grid population data from WorldPop and LandScan, the population data simulated by the method exhibited higher accuracy within the urban area and more accurately reflected the actual distribution of permanent residents.
Keywords:
100-m grid
building volume
spatialization
statistical model
urban permanent population
Journal
F
IF:
2
Papers:
404
Citations:
1.7W
Organization
Cited Papers