Detecting Urban form Using Remote Sensing: Spatiotemporal Research Gaps for Sustainable Environment and Human Health

Remote sensing offers large-scale and longitudinal assessment of the size, density, and function of cities associated with the sustainable environment and human well-being. In this chapter, we synthesize 376 peer-reviewed studies on urban land cover, building density, three-dimensional (3-D) structure, and land use using remote sensing approaches. We evaluated the sources of data, detection methods, as well as the spatiotemporal characteristics (e.g., locations and spatiotemporal scales). Our review identifies three research gaps: (1) Many urbanization studies monitor urban/non-urban change for a long period but not for the patterns of 3-D urban structure; (2) Increasing number of studies use deep learning approaches to detect urban land cover in large scales, especially with Sentinel-2, but there is a lack of time-series analysis and temporal accuracy assessment; (3) most of the urban land change studies focused on North America and East Asia but not in the Global South. A dilemma lies behind these research gaps: newer, high-resolution imagery, able to detect nuanced urban attributes, has a relatively short temporal span. In contrast, older imagery can detect long-term changes but has a lower resolution. This problem has led to a considerable paucity in investigating long-term urban dynamics. For instance, most of the studies investigating 3-D urban form covered less than five years. Dealing with these issues, recent developments in data fusion, temporal accuracy assessment, object-based image analysis, and deep learning methods are showing promise to enhance spatial resolution, extend temporal coverage, and to characterize land use intensity and 3-D structure, which are important factors affecting temperature, physical activities associated with public health. The increasing availability of computational power such as via Google Earth Engine allows analysis at large spatiotemporal scales such as comparing urban form and sustainable/health outcomes across multiple cities. We foresee increasing importance of remote sensing in providing evidence-based knowledge for policies and science of health cities.