Examining Urban Built-up Volume using High Spatial Resolution SAR and
Lidar Data: A Case Study in Detroit, Michigan, USA
Abstract
Accurate mapping of urban infrastructure, specifically buildings, and
extent is a high priority in addressing environmental and socioeconomic
problems. Importantly, such mapping must account for the
three-dimensionality (3D) of the urban environment that has
traditionally been lacking (e.g., land cover and land use
change—LCLUC—analyses often utilize two-dimensional satellite
imagery). Undoubtedly, considerable development and change in urban
areas takes place in the vertical dimension. Light detection and ranging
(lidar) data provide the ability to map in 3D, but systematic data
coverage both spatially and temporally is direly insufficient globally.
Previous works have successfully implemented spaceborne radar data such
as 1-km QuikSCAT data in lieu of lidar data to assess 3D urban build-up.
Higher spatial resolution is required, though, to examine urban
characteristics in detail. In this study, we test whether Sentinel-1
C-band Synthetic Aperture Radar (SAR) data can monitor 3D urban build-up
in high spatial resolution (i.e. 40-m) in Detroit, Michigan, USA, 2015.
Findings confirm, through comparison with aggregated 1-m lidar data, the
utility of high spatial resolution SAR data for examination of urban
built-up volume. Correlation and regression results show a strong linear
relationship between SAR backscatter and lidar volume citywide.
Moreover, SAR data, unlike coarser resolution radar data, can detect 3D
anomalies over time (e.g., new building construction, building
demolitions, etc.) further attesting to their utility for comprehensive
3D urban analyses a lower cost and higher repeatability over a much more
extensive coverage compared to lidar data.