Remote sensing for population estimates in small towns

Abstract

Studies on the dynamics of occupation and urban expansion impact the planning of public policies, as well as the allocation of resources for their implementation. As one of the UN's Sustainable Development Goals (SDGs) is to develop sustainable cities and communities, these studies gain relevance. Updating urban population data is important for securing financial resources to manage and plan cities. However, due to the complexity of such surveys, obtaining updated data is not trivial to support decision-making by urban managers. In this context, this study investigates Remote Sensing as an alternative for estimating urban population during inter-census periods. The objective was to verify the linear correlation between urban population and urbanized area in municipalities of the Nova Alta Paulista region, namely: Adamantina, Dracena, Junqueirópolis, Lucélia, and Osvaldo Cruz—for the years 2000, 2010, and 2022. The methodology involved collecting population data from IBGE and surveying the urbanized area using images from the Landsat satellite. Linear regression models were developed, and their accuracy was assessed using mean absolute error metrics. The results indicated a strong correlation between the growth of urbanized area and population increase; however, the mean error indicated that the model has limitations for population estimation, and testing additional variables is recommended. Finally, it was possible to project the population of these cities for the year 2030. The development of this research concludes that Remote Sensing is a supportive tool for population estimates, contributing to the planning and formulation of urban public policies.