Sensing Urban Physical Environment with GeoAI and Street-Level Imagery

Understanding and effectively managing urban spaces hinges upon the vital practice of sensing urban physical environments. It not only aids urban researchers in uncovering hidden complexities within the cityscape but also allows for a broader exploration of urban attributes that extend beyond mere physicality. In this age of information technology, a novel urban sensing paradigm has emerged, driven by geospatial artificial intelligence (GeoAI) and street-level imagery (SI), revolutionizing our ability to comprehensively perceive urban environments. Distinguished from conventional sources, SI offers a unique human-centric perspective. This viewpoint enables the extraction of extensive information, encompassing tangible elements within images, as well as intangible factors such as socio-economic characteristics, urban functionality, and even human perception—all analyzed through the prism of GeoAI. This chapter provides a comprehensive overview of research in urban physical environment sensing through the integration of GeoAI and SI. It delves into the existing imagery datasets, the GeoAI methodologies employed for their interpretation, and explores present applications while considering future trends. With the continuous advancement of robust AI models and the availability of high-quality SI, the amalgamation of GeoAI and SI holds great promise for amplifying urban sensing capabilities, enabling more profound understandings and effective administration of our cities.