FLOOD & MULTI-HAZARD EARLY WARNING
Home » Flood Control Solutions »
Early warning you can see.
WaveSave combines sensor data, automated visual monitoring and decision support to help authorities understand not only that conditions are changing, but what is changing, where it is happening and why it matters.
Book a demoVisual monitoring in action
A flood early warning system, also called a flood monitoring and warning system, gives people time to act before the water arrives. The UN initiative Early Warnings for All describes four parts that have to work together: knowledge of the flood risk; detection, monitoring and forecasting of rainfall and water levels; warnings that reach the right people in time; and the preparedness to respond. WaveSave focuses on monitoring and warning: sensors measure water levels and rainfall, cameras show what is changing on site, and operators get the information they need to decide sooner.
WHY VISUAL INTELLIGENCE MATTERS
A water-level or rainfall sensor provides a precise measurement at a specific location. Visual monitoring adds context across the surrounding area, helping show what may be changing between or around sensor locations.
Excellent for precise measurements at known locations.
Precise data — from the point being measured.
Adds context across the visible monitored area.
Context that helps explain what the measurements may mean.
The goal is not to replace sensors. It is to combine precise measurements with visual information so decision-makers have a clearer picture of what is happening.
HOW THE SYSTEM WORKS
An effective early warning system must do more than collect data. It must turn changing conditions into information people can understand and act on.
Collect the right information
Combine imagery, rainfall, water levels, weather and other relevant sensors.
Interpret what is changing
Use visual analysis, measurements, history and thresholds to identify meaningful change.
Put the event in context
Connect events to location, imagery, measurements and supporting information.
Trigger the appropriate response
Trigger inspection, escalation or operational response based on the risk.
The objective is not to generate more alerts. It is to give the right people enough information to make a better decision sooner.
FROM ALERT TO EVIDENCE
When monitored conditions trigger an event, operators should be able to understand what happened and investigate the evidence behind it. The iMaster/DocuCam platform is designed to connect warnings with the information needed to assess the situation quickly.
Example iMaster/DocuCam monitoring interface from the Rwanda deployment.
Depending on the monitoring configuration, an event can connect the warning to supporting information such as:
The result is a traceable record of what happened, where it happened and when.
A threshold tells you that something changed. Context helps your team decide whether that change needs inspection, escalation or immediate action.
FIELD EVIDENCE
The iMaster/DocuCam platform was installed in Rambura, Rwanda, as part of a government-supported pilot led by Rwanda’s Ministry in Charge of Emergency Management and Rwanda Space Agency, with GIZ funding and implementation by Hesotech. The results were later published in the peer-reviewed journal Progress in Disaster Science.
Example suspicious observation report from the Rambura, Rwanda deployment, showing before-and-after imagery linked to a detected event.
The Rwanda study reports continuous collection of high-resolution imagery and weather data across the monitoring period. The system remained operational during severe weather associated with the May 2023 disaster.
The significance is not simply that the system captured photographs. It turned visual change into measurable, traceable information.
The study documented visible changes that appeared before subsequent slope failure. In one case, a growing erosion feature at a road boundary developed over several days before a landslide. In another, cracks along terraces indicated approximately one day of potential pre-warning.
These examples do not mean that every landslide — or every flood — can be predicted days in advance. They show why systematically recording and comparing visual change can add valuable information to conventional monitoring.
Warning time and detectability depend on the hazard, site conditions, monitoring frequency, visibility and the indicators being monitored.
Read the peer-reviewed study →
Ba Town is a commercial hub in Fiji. Frequent flooding in the Ba River basin threatens critical infrastructure, disrupts market operations and endangers local communities. In 2023–2024, for Fiji’s Ministry of Waterways and with financial support from the Government of the Netherlands, WaveSave delivered an integrated flood management system in three parts:
Read the full case: Early warning system in Ba Town, Fiji
How visual monitoring works
DocuCam does not simply record continuous video. A pan-tilt-zoom camera systematically captures a defined area as a structured series of geo-referenced images, allowing the same location to be compared repeatedly over time.
The camera begins with an overview of the monitored area and progressively captures defined image tiles at the required zoom level. Each tile can be linked to a specific viewing position and geographic location.
Because the same part of the landscape is recorded repeatedly, images from different moments can be compared automatically. Visible changes can then be quantified rather than relying on someone manually watching a camera feed.
Computer-vision or other evaluation algorithms can analyse incoming imagery for relevant changes. The underlying image remains available, so operators can inspect the evidence behind a detected event and compare it with historical observations.
01
Capture
Record the same monitored locations systematically.
02
Compare
Measure how those locations change over time.
03
Interpret
Use defined algorithms and thresholds to identify relevant change.
This is what makes visual monitoring fundamentally different from ordinary CCTV: the imagery becomes structured data that can be searched, compared and analysed.
Visual monitoring can be combined with other relevant measurements and spatial information where the monitoring objective requires it.
Weather
Water levels
Ground movement
GIS & remote sensing
Existing systems & APIs
Designing a reliable monitoring system
The documented configuration supports low-power, solar-capable operation, mobile connectivity and local data buffering. Optical monitoring still depends on visibility, line of sight and a site-specific monitoring design.
Designed for field operation
Design limitations
Use visual monitoring where context matters. Use dedicated sensors where precise point, subsurface or continuous measurements are essential. Combine them when both are needed.
Applications
Different organisations need different information from an early warning system. The monitoring design should therefore start with the decision that needs to be made — which locations matter, which changes need to be detected and what action should follow.
For an example in an urban setting, see how WaveSave supports an early warning system in Ba Town, Fiji.
What is happening at the locations most likely to affect people and services?
Monitor flood-prone roads, urban waterways, culverts, drainage infrastructure, retention areas and other vulnerable locations where changing conditions may require early intervention.
What is happening around the measurements we already receive?
Combine water-level, rainfall and other sensor data with visual information around rivers, embankments, hydraulic structures and monitored catchment locations.
Is the asset itself — or the terrain around it — beginning to change?
Add monitoring around roads, railways, dams, reservoirs, power infrastructure, embankments, slopes and other critical assets where visible change can affect safety or continuity.
What happened, where did it happen and who needs to respond?
Bring event location, monitoring data, imagery and warning information into operational procedures so responsible teams can assess, escalate and respond more quickly.
What monitoring approach is practical, maintainable and useful locally?
Develop monitoring systems around local risks, infrastructure and response capacity, including locations where power, communications or technical resources may be constrained.
From warning to response
Detecting a potentially dangerous change is only the beginning. The real value of an early warning system is the time it creates to assess the situation, decide what matters and activate the appropriate response before impacts escalate.
A useful warning needs a predefined path from information to decision. That means agreeing in advance who receives the warning, how it is verified, when it is escalated and which operational response follows.
1. Assess
Review the event, location, measurements and available visual evidence.
2. Decide
Determine whether the situation requires observation, escalation or immediate intervention.
3. Mobilise
Notify the responsible teams and activate the relevant operational procedure.
4. Act
Implement the measure needed to reduce risk to people, infrastructure or services.
Depending on the risk
The response depends on the hazard. The important part is that the warning is linked to a decision that has already been thought through.
The WaveSave difference
WaveSave approaches early warning as part of a wider water and climate resilience strategy. Monitoring can be connected with flood modelling, operational procedures, infrastructure protection, training and — where appropriate — temporary flood-protection measures.
Frequently asked questions
The right early warning system depends on the hazard, site and decisions that need to follow. These are some of the questions we are most often asked about visual monitoring and sensor integration.
A flood early warning system monitors conditions that may indicate increasing flood risk and communicates relevant warnings so authorities, operators or communities have time to act. Depending on the location, it may combine rainfall, water levels, weather information, forecasting models, imagery and other monitoring technologies.
A water-level sensor provides a precise measurement at a specific location. Visual monitoring can add information about the surrounding area, such as changing water extent, erosion, slope movement or changes around infrastructure. The two approaches can complement each other rather than compete.
Yes, where the existing equipment provides a suitable electronic interface. The underlying platform is designed to receive information from different data sources through configurable drivers and APIs. The appropriate integration depends on the existing equipment, communications infrastructure and project requirements.
WaveSave delivers the system together with specialised partners: Hesotech for monitoring and, where needed, AeroSense for drone data and TAHMO for weather stations. Our hydrologists carry out the analyses and modelling, and we provide training together with our partners.
Yes. The supplied technology documentation demonstrates optical water-level measurement by measuring the visible water surface against a fixed reference point. Whether optical measurement or a conventional water-level sensor is preferable depends on the site, visibility and required monitoring frequency.
Image evaluation can use computer vision and AI-based methods to identify and quantify relevant changes. In the Rwanda study, a computer-vision algorithm was used to measure changes in imagery and support automated event generation. The platform is modular, so different evaluation methods can be applied according to the monitoring objective.
WaveSave therefore does not describe the system simply as “AI flood prediction”. The important question is what data is being evaluated, how the change is detected and what evidence supports the resulting warning.
There is no universal warning time. It depends on the hazard, site conditions, monitoring frequency, available data and how quickly the event develops. Some gradual changes may provide useful precursor information, while rapidly developing floods can allow much less time.
This is why WaveSave avoids promising a fixed number of warning hours or days without first understanding the location and the hazard being monitored.
Plan the right monitoring approach
Tell us about the hazard, location, existing monitoring infrastructure and the decisions your team needs to make. WaveSave can help assess which combination of visual monitoring, sensors, modelling and operational response measures makes sense for your situation.
No one-size-fits-all package. We start with the site, the risk and the decision the warning needs to support.
