VLM output
A rural two-lane road is surrounded by extensive brown floodwater covering the adjacent agricultural fields on both sides. Floodwater has reached the edges of the road and partially covered the road shoulders. Several cars, a van, and an orange utility truck are stopped on the raised section of the road, while several people and cyclists are standing on the wet pavement near the water's edge.
- Road Flooding DetectedYes
- Flood Water On RoadwayYes
- Flood SeveritySevere
- Water Depth On RoadShallow
- Road PassablePartially
- Stopped Vehicles Count5
- Pedestrians On RoadYes
- Weather ConditionOvercast
Customer challenges
Flood monitoring relies on a limited number of water-level stations and sensors, leaving large monitoring gaps between them
Flooding is mostly discovered through manual patrols or citizen reports, so detection lags behind reality
Manually watching every city camera 24/7 isn't feasible — the staffing cost is too high
Recession isn't tracked systematically, making it hard to tell normal drainage from a drainage-system malfunction
Both the city government and residents expect faster and faster flood-response times
Customer benefits
No new hardware needed — existing cameras upgrade directly into flood sensors, extending coverage to locations gauge stations can't reach
Proactive AI flood monitoring — flooding and recession are reported in real time, so status is always known
Dispatch, road-closure, and traffic-control decisions happen earlier
Full flood-lifecycle logging, with automatic alerts when recession looks abnormal
Key features
Standing Water Detection
Detects standing water on the road surface.
Flood Depth Classification
Classifies flood depth on affected roads.
Passability Assessment
Assesses whether the road remains passable.
Before & after
| Before | After — with Linker AI Nexa |
|---|---|
| Before Monitoring limited to water-level stations, with major gaps between them | After — with Linker AI Nexa Full AI camera coverage reaching locations gauges can't — 121x wider monitoring coverage than water-level stations alone |
| Before Relies on manual patrols and citizen reports | After — with Linker AI Nexa 24x7 proactive AI monitoring, cutting monitoring labor cost by roughly 90% |
| Before Flooding's impact on traffic discovered too late, slowing dispatch and road-closure decisions | After — with Linker AI Nexa Decisions made roughly 30 minutes earlier |
| Before No systematic record of recession | After — with Linker AI Nexa Recession time logged automatically, with alerts when drainage looks abnormal |