About This Project

Deploy an AI-powered monitoring system that uses high-resolution satellite imagery and ML to automatically detect informal construction and code violations in seismic zones, without requiring physical inspections. The system uses computer vision (convolutional neural networks) to identify new construction from satellite image time-series, cross-references building permits from municipal databases, and flags structures built without permits for priority inspection. For existing buildings, the system estimates seismic vulnerability from structural signatures visible in imagery (building height, footprint, materials) and creates a risk heatmap for inspectors. The system addresses the core problem identified in the seismic building codes challenge: informal construction sectors bypass permitting entirely and inspection systems are underfunded. By automating detection, the system multiplies the effective capacity of limited inspection forces by 10-20x, focusing human inspectors on the highest-risk structures rather than random sampling. The system has been conceptually validated by similar systems: Brazil uses satellite-based informal settlement detection (DICLE/INPE) for urban planning, and Turkey's 2023 earthquake aftermath showed that satellite imagery could identify 40,000+ damaged buildings within 72 hours.

Timeline