Using high resolution remote sensing image to help population estimation in small cities

This paper introduces how high resolution remotely sensed images and GIS data can help population estimation in small cities through a case study in the city of Waterloo. The foundation of this study lies on the relationship between living space and population. This relationship has been valid in recent studies and using Remote Sensing (RS) technologies for population estimation has gained much attention in RS studies [1]. However, the methods vary a lot due to different landscape and dwelling types. In this paper, a typical small city is chosen as a case study to exam the reliability of the proposed population estimation methods. First, two different classification methods are compared to generate residential area and dwelling count: rule-based feature extraction and sample-based feature extraction. Also, Geographic Information System (GIS) techniques are used to enhance the classification accuracy. Then, two population estimation models are compared: dwelling count model and residential area model. At last, results and discussion are included.