A simplified structure-from-motion photogrammetry approach for urban development analysis

Abstract

Despite the increasing availability of remote sensing imagery, spatial data of adequate spatio-temporal resolution for urban development analysis and urban planning remains inadequate in some places. Structure-from-Motion (SfM) photogrammetry provides a low-cost option to generate such data at the desired spatial resolution. This study thus presents a simplified approach for mapping and performing urban development analysis in a suburban area (Kuje) in Abuja, Federal Capital Territory, Nigeria. Here, we compared a Google Earth® historical imagery of 2005 (baseline data) with SfM-derived imagery of 2019 (recent data) to quantify the rate and magnitude of urban development in the area. The Ground Control Points (GCPs) survey achieved a Root Mean Square Error (RMSE) of 31.01 mm and 24.34 mm in planimetry and elevation respectively while the SFM survey achieved an RMSE of 21.11 mm, 25.35 mm and 96.06 mm in the X, Y & Z coordinate axes respectively, which are within the acceptable limit (<10 cm). The urban development analysis revealed that over the past 14 years, an average of 8 buildings were constructed per year. The area also witnessed an increase in the construction of roads and an associated decline in the green areas. In addition, the building height assessment revealed that land is not being used parsimoniously. The findings in this study demonstrate the viability of SfM photogrammetry in creating digital maps for small suburban areas and in performing urban development analysis. Furthermore, the results obtained can aid town planners and development authorities to make informed decisions about urban development issues in the area.