Research on AI-driven Research on Spatial Vitality Identification and Micro-renewal Mechanism of Commercial Districts in China and South Korea

Abstract

Amid rapid advances in artificial intelligence, regeneration of historic commercial districts is shifting from experience-based intervention to data-informed decision-making. Comparing Myeongdong business distric in Seoul and the ZhongLou business district in Xi’an, this study integrates spatial production theory, urban vitality theory, and micro-renewal into an AI-driven “spatial perception–vitality identification–renewal feedback” framework. Historical street-view imagery, POIs, field surveys, ArcGIS data, and consumer behavior data are used to reconstruct ground-floor commercial changes from 2008 to 2025. Computer vision, text classification, spatial statistics, and stage-based comparison assess business types, vacancy, street interfaces, accessibility, and activity intensity. Results show that multi-source data integration reveals street-segment relationships among spatial form, pedestrian activity, and consumer behavior. The ZhongLou district displays stronger node-based commercial agglomeration, while Myeongdong performs better in pedestrian continuity and public-space quality. Both face vitality decline under external shocks, functional homogenization, and uneven active frontages. Accordingly, the study proposes context-sensitive strategies for vacant-space reuse, functional diversification, pedestrian-interface repair, and adaptive monitoring.