Geodesic Labs is the culmination of a passion project focused on evacuation scenarios and logistics challenges.
We apply state of the art computer vision algorithms to extract detailed road networks from overhead imagery. We pair those road networks with rapid vehicle detection algorithms to inform traffic patterns, logistics, and a host of other challenges.
As of Summer 2022, we decided that open sourcing our codebase would have the greatest impact, so we welcome those interested to explore and build upon our efforts.
Road Networks
Extract roadways and intersections over city-scale areas directly from overhead imagery
Rapidly Localize Vehicles
Localize vehicles of multiple types, and place detected vehicles in the appropriate roadway or parking lot
Logistics / Evacuation
Compute optimal logistics paths, or evacuation routes in an emergency
Road Speed Limits
Deep learning architectures utilize road features to determine safe traversal speed for each roadway
Critical Nodes
Knowledge of both road network layout and vehicle location permits determination of critical locations in the area of interest.
Traffic Congestion
Estimate traffic congestion in real time based purely on imagery.