Urban Visual Intelligence: Perceiving Cities with AI and Street level Imagery

The last decades have witnessed an increasing penetration of digital technologies in the physical space This ushered in a series of radical changes in how we understand, conceive, design, and live the city More recently, the emergence of geospatial big data and advances in artificial intelligence have opened up new opportunities to sense urban dynamics and evaluate the processes and consequences of urbanization In this talk, we will focus on four key issues in perceiving the urban physical environment 1 How to observe urban physical environment on a large scale? 2 How to derive semantic information from street level imagery? 3 How to quantify the physical environment of a place? And 4 How to understand the interactions between the built environment, urban dynamics, and the socio economic environment at the human scale? We discuss these issues and approaches through a conceptual framework, framework,' Urban Visual Intelligence', that illustrates how new image data sources and deep learning techniques can reshape the way we perceive cities, enabling us to study their built environments and their interactions with human dynamics at different scales, and facilitate the development of sustainable smart cities.

Seminar Poster