Person overboard detection Drone Guide

Person overboard detection is an important maritime drone application because locating someone in the water can become extremely difficult within a very short period of time. Waves, darkness, weather, sea state, vessel movement and the relatively small visible profile of a person can all reduce the chance of detection from deck level.

Drones can add a fast aerial search layer around ships, ferries, offshore platforms, ports and coastal operations. Equipped with high-resolution RGB cameras, thermal sensors, AI person detection and powerful zoom, a drone can quickly search the surrounding water, identify possible targets and provide live coordinates to the rescue team.

The strongest use of drones is not to replace established man-overboard procedures, lifeboats, coastguard support or vessel manoeuvres. Instead, the drone provides additional situational awareness from above. It can search areas that are difficult to observe from the ship, maintain visual contact with a detected person and help rescue teams navigate towards the correct location.

For larger vessels and offshore operators, future systems could use permanently deployed drones that launch automatically following a confirmed man-overboard alarm. Combined with AI and vessel navigation data, this could reduce the time between the initial alert and aerial search.

What Is Person Overboard Drone Detection?

Person overboard drone detection uses unmanned aircraft to search for, identify and track a person believed to have entered the water from a vessel, offshore platform or coastal location.

The drone may be launched manually by trained crew or, in more advanced systems, from an automated docking station. Once airborne, it searches the surrounding sea using visual and thermal cameras while transmitting live video to the vessel or rescue team.

AI can assist by identifying human-shaped objects in the imagery and highlighting potential detections for immediate human review.

Why Person Overboard Incidents Are Difficult

A person in the sea presents a very small target compared with the size of the surrounding search area. Only the head, shoulders or part of a flotation device may remain visible above the water.

The vessel may also continue moving for some distance before the incident is confirmed. Wind, current and waves begin moving the person away from the original position immediately.

Every minute therefore matters. The faster rescuers can establish the person’s approximate location, the more focused the search can become.

Why Use a Drone?

A drone provides a higher viewing angle than personnel standing on the vessel. From above, it can search a much wider area and look behind waves or around parts of the vessel that may obstruct the crew’s view.

The aircraft can also move independently from the ship. While the vessel turns or prepares rescue equipment, the drone can remain near the suspected location.

This separation between search and rescue functions is extremely valuable. The drone searches and maintains situational awareness while the vessel or rescue craft concentrates on recovery.

Rapid Launch

Speed is one of the most important factors in a man-overboard response.

A drone should ideally be stored in a location where it can be launched quickly. A complicated setup process reduces much of the operational benefit.

For vessels that regularly operate in high-risk environments, dedicated ready-to-fly systems can be more valuable than a general-purpose drone stored deep inside the ship.

Automatic Launch

Future maritime drone systems may launch automatically when the vessel’s man-overboard alarm is activated.

The drone could receive the vessel’s position, heading and the estimated location where the person entered the water. It would then fly directly towards the initial search area while crew members begin established rescue procedures.

Human supervision should remain part of the operational framework, but automation can reduce response time.

Initial Search Position

The first search location is normally based on the estimated point where the person entered the water.

The drone can receive this position from the bridge or from integrated vessel systems.

If the person was detected by a camera, wearable beacon or man-overboard sensor, that information can improve the initial search area.

Accurate timing is important because current and wind begin moving the person immediately.

Search Patterns

Once the drone reaches the area, it can follow a structured search pattern rather than flying randomly.

The exact pattern depends on the accuracy of the initial location, sea state and surrounding conditions. Autonomous mission planning can help ensure that the area is covered systematically.

The drone should remain flexible enough for the operator to redirect it immediately when a possible target is identified.

Expanding Search

If the person is not found near the initial location, the search area can expand progressively.

The drone can cover rings or other defined patterns around the estimated position.

Current and wind information can also shift the search area in the expected drift direction.

This allows the search to become more intelligent than simply expanding equally in every direction.

AI Person Detection

AI person detection can analyse live drone imagery and identify shapes that may represent a human.

When the system finds a possible target, it can place a marker or bounding box around the object and alert the operator.

This can reduce the chance that a small person is overlooked within a large image.

Human confirmation remains essential because waves, debris and flotation equipment can create false detections.

AI Person Detection Over Water

Detecting people over water is more difficult than detecting them on land. Waves create constantly changing patterns, reflections can appear and disappear, and only a small portion of the person may be visible.

AI models therefore need training specifically for maritime conditions rather than relying only on general person-detection datasets.

Performance also depends strongly on altitude, camera resolution, sea state and the clothing or flotation equipment worn by the person.

Thermal Imaging

Thermal cameras can provide an additional search capability, particularly in low light or darkness.

A human body may create a thermal contrast with the surrounding water, but the effectiveness varies significantly with water temperature, clothing, immersion time and environmental conditions.

Thermal imaging should therefore complement RGB cameras rather than be treated as a guaranteed detection method.

Thermal Detection at Night

At night, thermal imaging can become particularly valuable because normal cameras may struggle to identify a person without artificial lighting.

The drone can scan the water for thermal differences and then use the RGB camera or searchlight to confirm the target.

The strength of the thermal signature may decrease as the person’s exposed body area reduces or cools.

This makes rapid deployment important.

RGB Cameras

High-resolution RGB cameras remain essential because they provide the detailed visual information needed to confirm a person overboard.

Colour imagery can identify lifejackets, clothing and rescue equipment.

In daylight and good visibility, RGB may provide stronger identification than thermal imaging.

A professional maritime search drone should ideally provide both.

Optical Zoom

Optical zoom allows the operator to examine a suspected target without immediately flying directly above it.

This can reduce unnecessary aircraft movement and help confirm whether an object is a person, buoy or floating debris.

High zoom requires excellent gimbal stabilization because even small aircraft movements become highly visible.

Wide-Angle Search Cameras

A wide-angle camera is useful during the initial search because it covers a larger portion of the water.

Once a possible target is identified, the system can switch to a narrower zoom view.

This combination of wide-area detection and detailed confirmation is particularly effective.

AI can monitor the wide-angle feed continuously while the operator uses zoom for verification.

Dual-Sensor Gimbals

A dual RGB and thermal gimbal is well suited to person-overboard operations.

Both sensors point towards approximately the same location, allowing the operator to switch rapidly between visible and infrared imagery.

The thermal camera may identify a possible target while the RGB camera provides visual confirmation.

This reduces the need to reposition the aircraft between sensors.

Searchlights

A gimbal-mounted searchlight can support night operations.

Once thermal or low-light imaging identifies a potential target, the light can illuminate the area for confirmation and help the rescue crew see the person.

The searchlight can also provide reassurance to the person in the water by showing that rescuers have located them.

Lighting should be managed carefully because reflection from the water can reduce camera visibility.

Low-Light Cameras

Modern low-light cameras can capture useful imagery in very limited illumination.

Moonlight, vessel lighting or distant port lights may provide enough light for image enhancement.

Low-light RGB complements thermal sensing because it provides more recognisable visual detail.

The best search payload may combine thermal, low-light and optical zoom.

Autonomous Target Tracking

Once a person is detected, the gimbal can automatically keep them centred within the image.

The aircraft may hover nearby or adjust its position as the person drifts.

This reduces operator workload during a stressful rescue situation.

The drone can then act as a persistent aerial marker above the casualty.

Maintaining Visual Contact

Maintaining contact after detection is often more important than continuing the wider search.

The drone can remain above or beside the person while transmitting their location to the vessel.

This helps prevent losing the target again in waves or darkness.

Battery management becomes critical because the aircraft may need to remain on station until a rescue craft arrives.

Target Geolocation

The drone can estimate the casualty’s coordinates using aircraft GNSS position, gimbal angle and, in some systems, laser range information.

These coordinates can be transmitted to the vessel or rescue craft.

Even an approximate position can be valuable when combined with live video.

Higher-quality navigation and gimbal metadata improve the accuracy of the estimated location.

Laser Rangefinders

A laser rangefinder may help determine the distance between the drone and the detected person.

Combined with aircraft position and gimbal orientation, this improves geolocation.

The technology can be useful when the aircraft remains at a safe distance.

Any laser system must be suitable for the intended maritime and aviation environment.

GPS Marker Generation

Once the person is confirmed, the drone software can create a persistent map marker at the detected location.

The marker can update as the person moves.

The rescue crew sees the latest position rather than relying only on the original man-overboard point.

This is particularly useful when current is moving the casualty rapidly.

Drift Tracking

A person in the water moves according to wind, waves and current.

The drone can observe this movement directly.

Software can calculate the direction and approximate speed of drift from successive positions.

This information can help the rescue team anticipate where the casualty will be by the time they reach them.

Current and Wind Integration

Advanced systems can combine observed drift with environmental data.

Wind speed, current direction and sea state can help estimate future position if visual contact is temporarily lost.

This should support rather than replace established search-and-rescue modelling.

Observed real-world drone tracking provides particularly valuable local information.

Vessel-Relative Navigation

Ships are themselves moving targets, making maritime drone operations more complicated than land-based search.

A drone may need to navigate relative to the vessel rather than only geographic coordinates.

Computer vision and vessel telemetry can help the aircraft understand where the ship is during launch, search and recovery.

This becomes particularly important for automated systems.

Moving Vessel Launch

Launching from a moving ship requires procedures different from normal ground operations.

The vessel may be pitching, rolling and changing direction.

The aircraft needs enough clearance to avoid superstructure and antennas.

For many operations, the vessel may reduce speed or adjust heading to support safe launch.

Moving Vessel Recovery

Landing on a moving vessel can be even more difficult than take-off.

The landing area moves in several axes and may be affected by strong local airflow.

Visual landing markers, RTK and relative-positioning cameras can help.

Some systems may use a net or other recovery method instead of conventional landing.

Automated Vessel Landing

Advanced maritime drones can potentially track a landing marker on the vessel and adjust continuously during descent.

The aircraft follows the ship rather than attempting to return to one fixed GNSS coordinate.

This could make Drone-in-a-Box systems practical on larger vessels.

Reliable automated recovery would be one of the key technologies enabling permanent maritime drone deployment.

Drone-in-a-Box at Sea

A shipboard Drone-in-a-Box system could remain protected and charged in a weather-resistant docking station.

Following a man-overboard alert, the system could launch quickly and begin searching around the vessel.

After the incident, it returns to its dock and recharges automatically.

Saltwater, vessel movement, wind and limited landing area make shipboard docking substantially more challenging than land-based installations.

Ferries

Passenger ferries are a strong potential application because they carry large numbers of people and operate repeatedly over defined routes.

A permanently installed drone could provide additional aerial situational awareness following a person-overboard alert.

The aircraft could also support other vessel inspections or emergency-response missions when not being used for rescue.

Operational procedures would need to integrate closely with the bridge team.

Cruise Ships

Large cruise ships can be hundreds of metres long and contain many areas from which a person could enter the water.

A drone provides an aerial search capability independent of the ship itself.

Thermal and optical zoom can be particularly valuable at night.

Integration with onboard CCTV or automated man-overboard detection systems could significantly reduce search initiation time.

Cargo Ships

Cargo vessels often operate with smaller crews and may be far from external rescue resources.

A drone could provide immediate aerial search capability following a man-overboard event.

Longer-endurance systems may be particularly useful because the vessel could be operating far offshore.

The drone can also serve routine hull, deck and infrastructure inspection purposes between emergency missions.

Offshore Platforms

Offshore oil, gas and renewable-energy installations are another strong application.

Personnel may work close to water and rescue resources can be limited by weather and distance.

A drone stationed on the platform can rapidly search the surrounding area following an alarm.

The same system may also support infrastructure inspection and security patrol.

Offshore Wind Farms

Technicians transferring between vessels and offshore wind turbines operate in challenging marine conditions.

A drone could support person-overboard response around turbines or service vessels.

The aircraft can provide an overhead view while rescue boats approach.

Wind conditions around turbines create additional flight-planning challenges.

Ports and Harbours

Ports can use drones for emergency water searches around docks, terminals and harbour infrastructure.

Person-overboard events in ports may involve complex structures, vessels and restricted visibility.

A drone can rapidly search between ships or around dock areas.

Coordination with port authorities and other aircraft operations is essential.

Coastal Search and Rescue

Drones can also support people reported missing in coastal waters.

The aircraft can search beaches, cliffs, rocks and nearshore areas while conventional rescue teams operate from land or sea.

Thermal imaging can improve low-light searching, although water conditions still affect performance.

Long-range coastal missions may benefit from fixed-wing VTOL aircraft.

Lifejacket Detection

AI can potentially identify high-visibility lifejackets more easily than a small human body.

Bright colours can provide useful visual contrast against dark water.

Computer vision models can therefore be trained to identify both people and flotation devices.

Detection becomes harder when the lifejacket is partially submerged or lighting is poor.

High-Visibility Clothing

High-visibility clothing can significantly improve aerial detection.

Orange, yellow and other contrasting colours can be more visible from above.

This highlights how personal safety equipment and drone detection systems can complement one another.

Future maritime PPE may increasingly consider aerial machine-vision visibility.

Wearable Beacons

Wearable man-overboard beacons can provide another important information source.

If a crew member falls into the water, the beacon can transmit an alert or location.

The drone can be sent directly towards that position rather than searching a large area from scratch.

AI vision then provides visual confirmation.

AIS Man-Overboard Devices

Some maritime personal safety devices can transmit an AIS man-overboard message.

A compatible vessel system can identify the casualty location.

A drone system could potentially use this coordinate as its first search waypoint.

The aerial camera then helps confirm and continuously track the person.

PLB Integration

Personal Locator Beacons can provide broader emergency location information depending on the device and rescue system.

The drone does not replace the beacon.

Instead, location information from the beacon can help focus the aerial search.

Combining radio location and optical verification creates a stronger response.

CCTV Integration

Large passenger vessels may already use extensive CCTV coverage.

If onboard video identifies where the person entered the water, this information can immediately define the initial drone search point.

AI could potentially connect CCTV detection with drone launch automatically.

This greatly reduces uncertainty about where to begin searching.

Automated Man-Overboard Detection Systems

Dedicated man-overboard systems may use cameras, radar, thermal sensors or other technologies to detect a person falling from a vessel.

These fixed sensors provide persistent detection, while the drone provides mobile search and tracking.

The two technologies are therefore complementary.

The fixed system detects the event, and the drone follows the casualty.

Radar Integration

Marine radar can provide situational awareness around the vessel, although detecting a person directly can be difficult depending on conditions and technology.

More specialised short-range radar systems may support man-overboard detection.

The drone adds visual confirmation.

Sensor fusion can reduce reliance on any single detection technology.

AI Sensor Fusion

Future systems may combine CCTV, wearable beacons, vessel position, radar and drone imagery in one rescue interface.

Each sensor contributes information about where the casualty may be.

AI can combine these inputs and calculate the most likely search area.

The drone then provides continuous confirmation and tracking.

Search Area Mapping

The drone’s ground control system can display the area already searched.

This reduces duplicate coverage and helps ensure systematic operations.

The map can also show the vessel track, initial man-overboard position and current detected casualty location.

This gives the bridge team a much clearer understanding of the incident.

Search Coverage

Drone altitude influences search coverage.

Flying higher allows a larger area to be observed but makes the person appear smaller in the image.

Flying lower improves detection detail but reduces coverage.

AI performance, camera resolution and sea conditions therefore need to influence mission altitude.

Ground Sampling Distance Over Water

The concept of ground sampling distance still matters even though the target is on the sea.

If one image pixel represents too much area, a person’s head may occupy only a handful of pixels.

Reliable AI detection requires sufficient target resolution.

Mission planners should therefore understand the practical person-detection range of the chosen camera.

Detection Range

Manufacturers may quote long camera or thermal detection ranges, but real-world performance depends heavily on sea state, weather and target size.

Detecting that something exists is different from confirming that it is a person.

Operators should therefore validate the system under representative maritime conditions.

Training exercises can provide valuable real-world performance data.

Sea State

Rough seas make person detection substantially more difficult.

Waves can repeatedly obscure the casualty.

Whitecaps may also create false visual or thermal patterns.

The drone’s own flight stability may degrade as wind increases.

Calm Water

Calm water provides much easier detection conditions.

A person or flotation device creates a clearer visual target.

Reflections can still create challenges depending on the sun angle.

Polarising optical approaches may help with some visible-light imaging, though they are not a complete solution.

Sun Glare

Sunlight reflecting from the sea can severely reduce camera visibility.

Changing drone position and camera angle can reduce glare.

Autonomous search algorithms could eventually optimise the viewing geometry.

Thermal cameras provide another sensor when RGB imagery is affected by strong reflection.

Fog

Fog reduces both human and camera visibility.

Thermal imaging may retain some capability depending on fog density and wavelength, but performance can still degrade significantly.

The drone should not be assumed capable of finding a person in all visibility conditions.

Flight safety may itself become the limiting factor.

Rain

Heavy rain reduces image quality and may exceed the aircraft’s weather limits.

Water on camera windows can make both RGB and thermal imagery difficult to use.

Person-overboard incidents do not necessarily occur in good weather, so maritime drones need realistic environmental qualification.

Operators should understand exactly what conditions the system can support.

Strong Wind

Wind affects aircraft endurance, stability and ability to return to the vessel.

A drone searching downwind may need substantial energy to fly back against the wind.

Energy planning must therefore consider the return journey continuously.

The strongest maritime rescue systems should calculate return energy dynamically.

Night Operations

Many person-overboard incidents may happen at night, particularly on vessels operating continuously.

Thermal imaging, low-light cameras and searchlights therefore become especially important.

Night operation also increases the challenge of launch and recovery.

Permanent maritime systems should be designed with night operations in mind rather than treating them as an optional extra.

Battery Endurance

Search missions can consume significant battery because the drone may need to fly quickly to the search area and then hover or perform repeated patterns.

Once the person is located, the aircraft may need to remain nearby until rescue is complete.

Battery reserves must also account for returning to a moving vessel.

A second drone or rapid battery replacement can provide useful redundancy.

Multi-Drone Search

Larger vessels could eventually use more than one search drone.

Two aircraft can cover separate areas simultaneously, substantially increasing search coverage.

Once one drone locates the casualty, the other may return or provide an additional observation angle.

Fleet coordination software would be needed to maintain safe separation.

Backup Drone

A backup aircraft can be particularly valuable if the primary drone reaches low battery while the person is still in the water.

The second drone can take over visual tracking.

This prevents losing aerial contact during battery replacement.

Large passenger or offshore vessels may find this redundancy worthwhile.

Hybrid Power Drones

Longer-endurance hybrid drones could provide extended search time, particularly for offshore operations.

However, they may be larger, noisier and more complex than battery aircraft.

For rapid shipboard launch, a compact electric multirotor may still be preferable.

The optimum platform depends on the vessel and operating environment.

VTOL Drones

Hybrid VTOL aircraft can provide longer range while still launching vertically from a vessel or offshore platform.

They may be useful when the search area expands significantly.

However, close hovering and target tracking can be more demanding than with a dedicated multirotor.

A mixed fleet may eventually provide both rapid close search and longer-range coverage.

Multirotor Drones

Multirotors are particularly suited to immediate person-overboard response because they can launch vertically, hover and reposition rapidly.

They can remain close to the casualty and provide a stable camera view.

The main limitation is endurance.

For most near-vessel incidents, however, their manoeuvrability is a major advantage.

Fixed-Wing Drones

Fixed-wing aircraft provide much greater search coverage for extended operations but cannot hover over the casualty.

They may therefore be more useful for wide-area search when the initial location is uncertain.

Once the casualty is found, a multirotor provides better persistent observation.

Platform selection should match the stage of the rescue mission.

Communications

The drone needs a reliable command and video link while operating around the vessel.

Large steel structures can block radio signals when the aircraft moves behind the ship.

Antenna placement on the vessel therefore matters.

Cellular or satellite connectivity may provide supplementary options depending on the operating area.

Direct RF Links

Direct radio links provide low-latency control and video over relatively short ranges.

For near-vessel person-overboard missions, this can be highly effective.

The antennas should be positioned to minimise blockage by the vessel superstructure.

Redundant control links can improve resilience.

4G and 5G

Near coastlines or ports, cellular networks may support drone telemetry and video.

However, offshore coverage may disappear rapidly.

A person-overboard system should not depend entirely on public cellular availability unless the operating area makes this reliable.

Direct vessel-based communication remains important.

Satellite Communications

Satellite systems can support longer-range offshore operations.

Because satellite bandwidth can be limited, the drone can perform AI processing onboard and transmit only essential video or detections.

For immediate near-vessel search, satellite may serve more as a backup or link to shore-based rescue coordination.

The local vessel-drone link still needs to remain robust.

Onboard AI

Running person detection directly on the drone reduces dependence on communications bandwidth.

The aircraft can analyse full-resolution video locally while transmitting only alerts and selected imagery.

This also reduces detection latency.

For emergency search, rapid onboard processing can be particularly valuable.

Edge Processing on the Vessel

The drone can transmit video to an onboard edge computer.

More powerful AI models can then analyse the imagery locally without requiring cloud connectivity.

The vessel retains control of sensitive data and receives immediate results.

This architecture is well suited to offshore operations.

Cloud Processing

Cloud AI is less suitable for the immediate detection loop because network latency and connectivity may be unreliable at sea.

It can still support post-mission analysis, fleet model improvement and training.

Operational detection should ideally remain local.

Emergency response cannot depend on a distant cloud connection.

AI False Positives

Waves, buoys, debris and seabirds can all resemble a person in aerial imagery.

AI may therefore generate false alerts.

The system should display the underlying image to the human operator for rapid confirmation.

A false positive is less harmful than missing a casualty, but excessive false alerts can distract rescuers.

AI False Negatives

A person may also be missed, particularly if they are partially submerged, obscured by waves or too small in the image.

AI confidence should therefore never be interpreted as proof that the search area is clear.

Human review, repeated passes and multiple sensors remain important.

Search-and-rescue procedures should not rely solely on AI.

Person Segmentation

Advanced AI can identify the exact pixels associated with the detected person rather than only drawing a bounding box.

This may improve tracking against complex wave backgrounds.

The segmentation can also help estimate how much of the person remains visible.

Such technology may improve future maritime AI systems.

Flotation Device Detection

AI can detect lifebuoys, life rafts or other flotation devices.

These may indicate the casualty’s location even when the person is difficult to see.

The system could therefore search for several classes simultaneously.

This broadens the search beyond a conventional person-detection model.

Object Tracking Through Waves

The casualty may temporarily disappear behind a wave.

Tracking algorithms can estimate where the person should reappear based on previous movement.

This helps the system maintain target identity through short periods of occlusion.

Longer periods without visual contact should trigger a renewed local search.

GPS Drift Marker

The system can maintain a virtual drift marker based on the casualty’s last known positions.

If visual contact is lost, the drone searches around the predicted location.

This is much more effective than returning immediately to the original man-overboard position.

The prediction can update using wind and current data.

Rescue Boat Guidance

Once the casualty is located, the drone can provide coordinates and a live overhead view to a rescue boat.

The aerial perspective can help the crew understand the casualty’s position relative to waves and the vessel.

The drone may also remain slightly ahead of the rescue craft to guide it towards the target.

Communications should remain simple and reliable during the emergency.

Lifebuoy Delivery

Some specialist drones can carry lightweight flotation devices or small rescue payloads.

In certain situations, a drone may be able to release a flotation aid near the casualty before the rescue craft arrives.

Payload weight, release accuracy and aviation rules need careful consideration.

The primary drone role should remain detection and tracking unless the system is specifically designed and approved for delivery.

Emergency Flotation Payloads

A compact inflatable flotation device could potentially be carried under a rescue drone.

Once the person is located, the device is released nearby.

This may provide additional support during the time before physical recovery.

The solution needs to be extremely reliable because inaccurate deployment could create false confidence.

Loudspeaker Support

A drone-mounted loudspeaker may allow rescuers to communicate with a conscious person in the water.

Instructions can be given while the rescue craft approaches.

The drone can also reassure the casualty that they have been located.

Wind and rotor noise limit intelligibility, so this should be treated as supplementary.

Searchlight as Visual Marker

A searchlight can serve not only for illumination but also as a marker for the rescue crew.

At night, the drone hovering above the person provides a highly visible reference point.

This can make the casualty easier for the rescue boat to locate.

The aircraft should maintain enough height and separation to avoid disturbing the person unnecessarily.

Vessel Bridge Integration

The strongest system integrates the drone directly with the vessel’s bridge systems.

The bridge can see the drone position, search area, detected casualty and live video on one interface.

The drone receives vessel heading, speed and navigation information.

This reduces separate devices and communications during an emergency.

Electronic Chart Integration

Person-overboard coordinates and drone position can potentially be displayed on the vessel’s electronic navigation systems or a linked emergency-response interface.

This provides geographic context for the bridge team.

The system can show how the casualty is drifting relative to the vessel.

Integration should be designed carefully to avoid interfering with certified navigation systems.

Automatic Incident Recording

The drone system can automatically preserve video, coordinates and timestamps from the search.

This creates a complete incident record.

Such data can support later safety investigation and training.

Access should be controlled because person-overboard incidents involve sensitive information.

Training Exercises

Regular exercises are essential before relying on a drone during a real emergency.

Crew members need to understand launch procedures, aircraft limitations and how AI detections appear.

Training also allows the operator to determine realistic detection ranges under different sea conditions.

Dummy targets can be used to test the complete workflow.

System Readiness

A rescue drone provides little value if its battery is flat, software is outdated or the aircraft is inaccessible when the alarm occurs.

The system should therefore have routine readiness checks.

Battery health, cameras, communications and positioning should be verified automatically where possible.

Drone-in-a-Box systems can support this continuous readiness.

Automated Pre-Flight Checks

A permanently deployed drone can monitor battery condition, GNSS, gimbal, cameras and communication links.

If one subsystem fails, the bridge receives a warning before an emergency occurs.

This is much better than discovering the problem during a real rescue.

Safety-critical drone applications require strong health monitoring.

Weather Monitoring

The system should know whether current conditions are within the aircraft’s limits.

Wind and precipitation can make flight impossible even during a serious emergency.

A rescue plan should therefore never assume that the drone will always be available.

It remains one layer within a wider person-overboard response.

Saltwater Protection

Maritime drones are exposed continuously to salt, spray and humidity.

These conditions accelerate corrosion of motors, connectors and airframes.

The aircraft should be designed or maintained specifically for marine environments.

Regular freshwater cleaning may be required depending on the system.

Waterproofing

Improved weather protection can increase operational availability.

However, waterproofing alone does not mean the aircraft can safely fly in heavy rain, extreme wind or breaking waves.

The complete aircraft, camera and dock need suitable environmental specifications.

Operators should work from defined limits rather than general claims of weather resistance.

Flotation for the Drone

Some maritime drones may use emergency flotation systems.

If the aircraft makes a controlled landing on the water or falls in, flotation can improve recovery chances.

This protects the equipment but does not necessarily make the drone capable of taking off again.

Aircraft recovery can be particularly important when operating far offshore.

Parachutes

Parachutes are commonly considered for reducing ground impact, but their value over open water is different.

A parachute may slow descent but can also move the aircraft farther from the vessel in strong wind.

For maritime rescue drones, flotation and robust propulsion redundancy may sometimes provide more relevant mitigation.

The correct solution depends on aircraft architecture.

Redundant Propulsion

A drone dedicated to emergency maritime operations may benefit from propulsion redundancy.

Larger multirotors with more than four motors can sometimes tolerate certain motor or propeller failures better than basic quadcopters.

The complete aircraft design needs to determine actual fault tolerance.

Redundancy adds weight but may be worthwhile for high-value emergency applications.

Battery Redundancy

Multiple battery systems can reduce the chance that one electrical failure causes complete aircraft loss.

They can also provide additional reserve.

The trade-off is increased aircraft mass.

For rescue operations, reliability may take priority over maximum endurance.

Cybersecurity

A person-overboard system connected to vessel networks needs appropriate cybersecurity.

Drone control, navigation information and video feeds should be protected from unauthorised access.

Software updates and user permissions require careful management.

Emergency systems need to remain secure without becoming unnecessarily difficult to operate.

Privacy

Person-overboard operations involve highly sensitive imagery.

Video may show injured or distressed individuals.

Access and retention should therefore be tightly controlled.

The system should collect and preserve only what is necessary for rescue, safety and legitimate investigation purposes.

Regulatory Considerations

Maritime drone operations remain subject to aviation rules even when launched from ships.

Requirements may vary according to airspace, distance from shore, aircraft type and operational model.

BVLOS, automated flight and night operations may require additional approvals.

Emergency-use procedures should be established before an incident rather than improvised during one.

International Operations

Commercial vessels operate across multiple jurisdictions.

A shipboard drone programme therefore needs to consider how aviation rules change as the vessel moves between countries and waters.

This can complicate permanent automated operations.

Large international fleets may need procedures defining where and under what conditions the drone can legally operate.

Search and Rescue Coordination

If coastguard, helicopter or other rescue aircraft become involved, the drone must not create an airspace conflict.

The vessel should communicate clearly that a drone is operating in the area.

The aircraft may need to land when crewed rescue aviation approaches.

Human life takes priority over maintaining the drone search.

Benefits of Person Overboard Drones

The biggest benefit is speed of aerial situational awareness. A drone can quickly move to the estimated incident location while the vessel begins its rescue manoeuvre.

Once the person is found, the drone can maintain continuous observation and provide updated coordinates. Thermal cameras and AI can support detection, particularly when visibility is poor.

The aircraft can also guide rescue craft and provide the bridge with a much clearer understanding of the casualty’s movement relative to the ship.

Faster Detection

Every minute lost increases the search area.

A rapidly launched drone reduces the time before an aerial search begins.

If integrated with automated man-overboard detection, the aircraft could be moving towards the casualty within moments of the alarm.

This is potentially one of the most important benefits of automation.

Better Situational Awareness

From deck level, crew members see the water from a relatively low angle.

The drone provides a top-down view that can show both the casualty and rescue craft simultaneously.

This makes it easier to understand relative position and movement.

The same live feed can be shared with several decision-makers.

Reduced Search Area

If the drone finds the person quickly, rescuers no longer need to search a large expanding area.

Resources can concentrate on recovery.

This is particularly valuable offshore where external search-and-rescue assets may take significant time to arrive.

Rapid localization can therefore directly improve rescue efficiency.

Continuous Tracking

Finding the person once is not enough if they disappear from view again.

The drone can hover nearby and track drift.

AI target tracking reduces operator workload while live coordinates continue updating.

This persistent aerial marker is one of the strongest capabilities a drone can provide.

Challenges and Limitations

Person-overboard drone detection faces significant challenges. Rough seas, strong wind, rain, darkness and fog can all make both flight and detection difficult.

A person may occupy only a very small number of pixels and can disappear repeatedly behind waves. Thermal signatures may also weaken depending on water temperature and immersion conditions.

Battery endurance, communications and moving-vessel recovery create additional operational complexity.

For these reasons, drones should form one part of a layered man-overboard response rather than replace established maritime rescue procedures.

The Future of Person Overboard Detection

Person-overboard drone systems are likely to become increasingly integrated with vessel safety systems rather than operating as standalone aircraft.

Future vessels may use CCTV, radar, wearable beacons and AI to detect the moment someone enters the water. The event would immediately create a geographic search point and trigger the shipboard drone system.

A drone could launch automatically from a protected deck station while the bridge begins the vessel’s standard man-overboard manoeuvre. Onboard AI would search the water using RGB and thermal imagery and compare detections from both sensors.

Once a person is identified, the aircraft would switch automatically from search mode to tracking mode. Their coordinates would update continuously as wind and current move them away from the original incident point.

A second drone could take over if the first aircraft reaches low battery, ensuring uninterrupted aerial observation until recovery is complete.

AI models will also become increasingly specialised for maritime detection. Instead of looking only for a complete human shape, systems will recognise heads, arms, lifejackets, flotation devices and partially submerged people.

Wearable devices could become part of the same ecosystem. A crew member’s beacon provides the approximate location while the drone supplies visual confirmation and continuous tracking.

Autonomous shipboard landing will be another important development. Once drones can reliably launch and recover from moving vessels, they can remain permanently available without requiring a specialist drone team onboard.

The biggest transition will therefore be from manually launching a drone after someone is reported missing towards integrated automated man-overboard response, where detection, aerial search, tracking and rescue coordination begin within the same connected safety system.

Conclusion

Person overboard detection is a compelling maritime drone application because one of the greatest challenges in a man-overboard incident is locating and continuously tracking a very small target in a large and constantly moving environment.

Drones provide a rapid aerial viewpoint that can complement bridge lookouts, CCTV, rescue boats, wearable beacons and established man-overboard procedures. High-resolution RGB cameras provide visual confirmation, while thermal sensors, low-light cameras and searchlights can extend capability into more difficult lighting conditions.

Artificial intelligence can assist by identifying possible people or flotation devices within the imagery and maintaining target tracking after detection. Accurate drone positioning can then provide continuously updated coordinates to the rescue team.

The greatest opportunity lies in automation. A permanently deployed maritime drone could launch immediately after a confirmed alert, travel to the estimated incident location and begin a structured search while the vessel prepares for recovery.

Drones cannot guarantee detection in every condition. Rough seas, extreme weather, poor visibility and aircraft limitations can all reduce effectiveness, and established maritime search-and-rescue procedures remain essential.

Their role is to provide an additional, fast and mobile aerial search layer.

For passenger vessels, commercial ships, offshore platforms, ports and maritime rescue organisations, combining drones with AI, thermal imaging, wearable beacons and automated detection systems could significantly improve the speed and quality of person-overboard response and help rescue teams maintain the one thing that becomes most difficult after an incident: continuous awareness of exactly where the person is.