Detecting urban functional zones changes via multi-source temporal fusion of street view and remote sensing imagery

• A Multi-Source Temporal Fusion Model is proposed for UFZs change detection. • A framework that fuses RSI and SVI is introduced for direct urban dynamic monitoring. • Experiments in three cities confirm the model and reveal UFZs transitions since 2014. • Statistical analysis reveals respective strengths of RSI and SVI across scenarios. Monitoring changes in urban functional zones (UFZs) is vital for optimizing land use, improving urban management, and promoting sustainable development, as these changes reflect critical socio-economic dynamics. However, due to the complexity and diversity of UFZs, existing methods face challenges in effectively capturing these changes. Remote sensing (RS) imagery offers top-down views but often misses ground-level details, while street view (SV) imagery provides close-range perspectives. However, effectively integrating SV and RS data for comprehensive UFZs change detection remains a challenge. This paper presents a deep learning-based method that leverages SV and RS data to enhance UFZs change detection. The proposed method employs an attention mechanism to dynamically fuse RS and SV features along with temporal change information. Over 100,000 SV images were classified, and their spatial distribution frequencies were used as features in the Multi-Source Temporal Fusion Model (MSTFM). MSTFM integrates RS and SV data via the Multi-Source Fusion Module (MSFM) and incorporates temporal features through the Multi-Temporal Fusion Module (MTFM), enabling classification and change detection. The method was applied in a large-scale case study across Datong, Chengdu, and Nanjing, achieving 86 % overall accuracy and 82 % Kappa for single-period UFZs classification, and over 93 % accuracy for binary change detection. Experimental results demonstrate the model’s effectiveness in integrating SV and RS imagery and capturing UFZs dynamics, with strong potential for identifying urban renewal and supporting sustainable development.