As urbanization continues to accelerate, the structural and functional complexity of cities grows, making the effective management and understanding of urban spaces increasingly critical. At the same time, advancements in the collection, sharing, and analysis have greatly improved our capacity to comprehend the impact of urban physical environments on human activities. Urban sensing technology, in particular, has emerged as an indispensable tool for studying urban dynamics and enhancing the quality of human life (Liu et al., Annals of the Association of American Geographers, 105:512–530, 2015). Computer vision technology has become increasingly important for understanding and representing the urban physical environment, making it a key tool for urban sensing (Zhang et al., Annals of the American Association of Geographers, 114:876–897, 2024). However, traditional computer vision methods often focus on technical improvement and data analysis, overlooking the most essential component of cities—people. Human-centered computer vision bridges this gap by not only analyzing physical spaces, but also focusing on residents’ behavior patterns, social interactions, and living needs (Luca et al., ACM Computing Surveys, 55(1):1–7:44, 2021; Salazar-Miranda et al., Landscape and Urban Planning, 234:104715, 2023). By gaining a deeper understanding of human-environment interactions, we can better investigate urban dynamics and formulate more sustainable development policies. This chapter discusses the role and applications of human-centered computer vision in urban sensing, focusing on how this technology can be used to enhance our understanding of city residents and improve their quality of life.