Abstract
Population data are key indicators of policymaking, public health, and land use in urban and ecological systems; however, traditional censuses are time-consuming, expensive, and laborious. This study proposes a method of modelling population density estimations based on remote sensing data in Hefei. Four models with impervious surface (IS), night light (NTL), and point of interest (POI) data as independent variables are constructed at the township scale, and the optimal model was applied to pixels to obtain a finer population density distribution. The results show that: (1) impervious surface (IS) data can be effectively extracted by the linear spectral mixture analysis (LSMA) method; (2) there is a high potential of the multi-variable model to estimate the population density, with an adjusted R2 of 0.832, and mean absolute error (MAE) of 0.420 from 10-fold cross validation recorded; (3) downscaling the predicted population density from the township scale to pixels using the multi-variable stepwise regression model achieves a more refined population density distribution. This study provides a promising method for the rapid and effective prediction of population data in interval years, and data support for urban planning and population management.
Highlights
Population data are considered important indices for the development of a country or region
After calculating th proportion of each endmember type in the external square of each point, the proportio of the impervious surface (IS) in the square was compared with the IS abundance value extracted by linear spectral mixture analysis (LSMA) Subsequently, the experimental and true data were compared by calculating the linea fitting between them
Our study explored the population density estimation model in Hefei, emphasising the importance of population data for some policies, such as the introduction of talent and household registration for migrant workers, which is closely concerned with the population
Summary
Population data are considered important indices for the development of a country or region. Land-use types and building areas can be extracted from remote sensing images for population estimation [15,16,17]. Since the methods for extracting ISs may lead to the confusion of ground objects [29], it is easy to lead to the estimation error of delineating population distributions with IS data as the sole independent variable [4]. In this study, the geo-spatiotemporal big data POI and the latest highresolution NTL data from Luojia-1 were fused with the IS information extracted from remote sensing images to try to establish the population estimation model. The study area was masked from the image using administrastepwise regression model for population estimation was applied to provide a reference for mapping the population density distribution
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
