Identify different types of urban renewal implementations at the streetscape scale

Understanding the different types of urban renewal implementation processes can inform ways to improve residents’ quality of life and optimize sustainable development strategies. Existing research has primarily focused on detecting pixel-level or object-level changes in urban physical space, but it frequently overlooks the semantic complexity inherent in urban renewal. This complexity involves an integration of what changed, where the change occurred, and how it occurred, and is important to distinguish the different types of renewal. To address this gap, this study provides a multi-type urban renewal identification (MTURI) framework using street view images (SVIs) to combine multi-source information at the street level. We constructed an SVI benchmark dataset and developed a comprehensive indicator system comprising built features, semantic attributes and street-level environmental factors. Using a two-stage recognition algorithm, the MTURI model identifies various types of urban renewal activities. We also applied the model to four renewal community cases to evaluate the potential applications of the method. The findings demonstrate that our model can effectively identify various types of renewal and provide evidence-based assessments of the effectiveness and benefits of urban policy interventions. Our paradigm introduces new tools and insights into research and practice in SVI identification.