The Center for Data Innovation recently spoke with Julia Tan, CEO of Geospan, an Australia-based company developing AI-powered road-intelligence tools. Tan explained how Geospan uses existing roadside cameras to detect hazards, turn them into real-time data, and share warnings with drivers and road operators.
David Kertai: What problem is Geospan solving?
Julia Tan: Many road hazards are detected too late. Floods, debris, wildlife crossings, and stopped vehicles can appear suddenly and fall outside the field of view of a vehicle’s cameras and other sensors. Existing roadside cameras may capture these events, but these traditional systems do not consistently turn what they see into real-time information that drivers can use. As a result, drivers may not know about hazards until they encounter them, causing delays, disruptions, and safety risks.
Geospan addresses this gap by turning existing roadside cameras into a live source of road intelligence. Our Pulse device is a compact computing unit that attaches to existing roadside cameras and uses an AI-powered computer-vision system to analyze videos for hazards and changing road conditions. Those detections flow into our cloud platform, Atlas, which organizes and distributes the information so fleets, road operators, maps, and other connected systems can respond. Instead of relying only on what a vehicle can see, Geospan gives it awareness of conditions farther along the road.
Kertai: How does the Pulse device detect hazards and changing road conditions?
Tan: Pulse connects to existing roadside cameras and processes their video in real time. It runs an AI system that analyzes camera footage locally at the roadside, identifying hazards and changes in the road environment as they occur. This allows operators to add Geospan’s detection capabilities to existing camera infrastructure without replacing the cameras themselves.
The system can also track how conditions change over time. For example, it can determine whether floodwater is rising or receding, whether debris continues to block a lane, or whether a stopped vehicle has started moving again. By continuously analyzing the same area, Pulse turns a camera from a device that simply records video into a system that identifies and tracks changes on the road.
Kertai: How does Geospan turn detected hazards into real-time warnings for drivers and vehicles?
Tan: Pulse sends its detections to our cloud platform, Atlas, which timestamps each event and identifies its precise location on the road, including the relevant road, lane, or zone. Atlas then makes that information available through application program interfaces that are connected to maps, fleet platforms, vehicle systems, and other infrastructure so they can deliver the appropriate warnings.
Atlas also tracks each detection as an ongoing event. As conditions change, the system can update an event, escalate or reduce its status, or close it when the hazard disappears. For example, once debris has been cleared from a lane, Atlas can update the event so connected systems no longer treat it as an active hazard. This creates a continuous link between what roadside cameras detect and how organisations respond.
Kertai: Could you share any use cases that demonstrate how organizations are using Geospan?
Tan: One example is wildlife-collision prevention. We are working with local governments in areas with high wildlife-collision rates, where Geospan provides information about when and where animals enter the road. Operators can use this information to identify areas where wildlife frequently crosses, evaluate how existing mitigation measures are performing, and understand how wildlife activity affects traffic. For example, they can use repeated detections to identify periods when animals are most active and determine where additional warnings or other safety measures could help reduce collisions.
The same detections can also provide fleets with more timely information about wildlife on or near the road. Instead of relying on a driver to spot an animal, connected systems can use Geospan’s data to alert drivers to a potential hazard before they reach the area. This gives drivers more time to slow down or take other appropriate action.
Kertai: What is your long-term vision for Geospan and road intelligence?
Tan: Transportation is becoming more connected and increasingly automated, but vehicles still depend on information about the road environment beyond what their own sensors can see. Our long-term vision is for Geospan’s roadside vision layer to extend a vehicle’s awareness beyond its direct line of sight by providing real-time information about hazards and changing road conditions farther ahead.
As vehicles, maps, and road infrastructure become more connected, we want Geospan to provide the live road intelligence that allows these systems to respond to changing conditions. This could help drivers, fleets, and road operators make faster, better-informed decisions about what is happening ahead.
