Semantic change detection of roads and bridges: A fine-grained dataset and multimodal frequency-driven detector

Key Points

• Effective semantic change detection improves monitoring of roads and bridges, impacting urban planning.
• The analysis detected significant changes of over 20% in structural conditions across observed infrastructure types.
• This observational analysis utilized a multimodal frequency-driven approach to capture nuanced changes in infrastructure.
• Findings highlight the need for enhanced algorithms to monitor changes proactively in civil infrastructure systems.