WHU-Infra3D: A Full-stack Multi-modal Dataset and Benchmark for 3D Roadside Infrastructure Inventory
Researchers have introduced WHU-Infra3D, a new large-scale, multi-modal dataset designed for the precise digitization of urban infrastructure. This dataset covers 53.8 km across three cities, integrating panoramic imagery and LiDAR point clouds with rigorous 2D-3D instance association and cross-frame tracking. It includes over 175,000 2D bounding boxes and thousands of 3D infrastructure instances, along with detailed attribute and status annotations to aid in operational health assessment. The dataset aims to serve as a testbed for advancing AI-driven urban infrastructure inventory and lifecycle management, highlighting existing domain gaps and model vulnerabilities. AI
IMPACT This dataset aims to improve AI's ability to inventory and manage urban infrastructure, potentially leading to more efficient maintenance and digital twin city development.