AI-enhanced Real-time 3D Coastal Reconstruction for Enhancing Resilient Communities in Southern Lake Michigan

Major Goals and Objectives

To develop an AI-enhanced 3D reconstruction workflow that integrates UAV imagery with existing aerial and satellite data to generate high-resolution, real-time georeferenced models of coastal and watershed features in the southern Lake Michigan region. To apply this system to monitor and quantify environmental changes—such as shoreline erosion, dune morphology, stormwater runoff, and infrastructure vulnerability—before, during, and after extreme weather events or seasonal transitions. To evaluate the performance and accuracy of state-of-the-art reconstruction methods (VGGT, MASt3R, DUSt3R) for coastal applications, using ground-truth data (e.g., GNSS, LiDAR) to validate outputs and assess model limitations under varying conditions. To create an open-access toolkit and decision-support platform—including a web-based dashboard and immersive VR/MR interface—that enables stakeholders to visualize 3D results and extract actionable metrics (e.g., erosion rates, flood extent, asset risk). To engage with community stakeholders and IISG outreach specialists from project inception to ensure research findings are translated into practice through training workshops, user guides, and integration with local planning and public outreach efforts.

Research Information

Principal Investigator:
Wei Wu
Initiation Date:
2025
Affiliation:
Purdue University

Contacts

Qingchun Li
qingchun@purdue.edu
Skip to content