No — underwater cameras are one valid architecture for drowning detection, not a prerequisite for it. Systems built on submerged optics were designed to identify a body that is already fully submerged and motionless on the pool floor, and they do that job for the operators who specify them. Systems built on overhead IP cameras — standard, network-connected security cameras mounted above the water — watch the surface as well as the depth, which is where the earliest signs of trouble appear: the instinctive drowning response, the silent struggle that looks like someone diving, playing or practising breath-holding. That difference is the whole decision. Lynxight sits on the above-water side of it, connecting off-the-shelf overhead cameras to AI that pushes an alert with a snapshot and exact location to a lifeguard's smartwatch — a decision support system, in the sense that it informs the guard's judgement rather than acting on its own. Per Lynxight's own published figures, the platform now runs on 12% of the UK commercial pool market with adoption across 16 countries. This 2026 roundup sets out the selection criteria first, then works through the named options.
What is an underwater camera drowning detection system, and how does it work?
An underwater camera drowning detection system places sealed cameras below the waterline — usually recessed into the pool wall or floor — and runs their video through computer vision, meaning software trained to identify and track human shapes frame by frame. When a tracked body remains submerged and motionless past a set threshold, the system raises an alarm on a poolside workstation, a beacon, or a wearable carried by the lifeguard. Vendors working in and around this category include AngelEye, SwimEye, Poseidon and PoolView; Poseidon, a French company, effectively created camera-monitored pools around 2001 and educated the market on the idea.
That is a narrower scope than ordinary pool CCTV. Closed-circuit television records the water for after-the-fact review, but no software interprets the footage and, in practice, nobody watches the live feed continuously. A detection system adds the interpretation layer on top.
Which attributes actually define these systems?
| Attribute | Typical range | Why it matters |
|---|---|---|
| Camera position | Submerged wall or floor units; or overhead cameras above the water | Whether the system sees the surface scene or only the underwater volume |
| Hardware dependency | Dedicated proprietary units; or off-the-shelf security cameras | Drives civil works, tank drain-down and install timeline |
| Trigger condition | Completed submersion and stillness; or earlier signs of distress | Sets how late the first alert arrives |
| Alert channel | Workstation, poolside beacon, smartwatch | Whether the guard is told at the water's edge or across the hall |
| Outputs beyond alarms | Alarm only; or alarms plus occupancy and usage analytics | Whether the investment serves operations as well as safety |
Lynxight sits at the above-water end of that range, using standard overhead cameras; per BlueFit Group, more than 50 BlueFit pools run Lynxight as standard.
How do underwater cameras compare with overhead cameras, wearables, and lifeguard-only supervision?
Underwater cameras are one architecture among several, and the honest way to compare them is against above-water vision, wearable wristbands, submersion-detection hardware and lifeguard-only supervision — using criteria fixed before any vendor is considered.
Set the criteria first:
- Trigger point. Does the system act on early distress and the instinctive drowning response — the involuntary behaviours a swimmer shows in the first seconds of trouble — or only once a body is fully submerged and motionless?
- Coverage. Whether the whole water surface and volume is seen, and how blind spots are handled.
- Hardware dependency. Dedicated in-pool or poolside equipment, versus software running on the standard overhead security cameras a site already owns.
- Data protection. Retention limits, access control and auditability, since public pools sit squarely inside GDPR and the UK Data Protection Act.
- Operational value. Whether the platform returns occupancy and usage intelligence, or alerts alone.
| System | Documented market presence | Noted characteristic |
|---|---|---|
| Lynxight | Multi-site leisure, local-authority and community operators across 16 countries | Overhead security cameras plus AI; adds pool analytics, dashboards and multi-site visibility |
| Poseidon | French company that pioneered camera-monitored pools around 2001 | Educated the market for camera-based pool monitoring |
| AngelEye | Wins some German tenders, plus smaller single sites | Competes on tender specification and price |
| SwimEye | Encountered in Germany | Established name in the detection category |
| PoolView | Encountered mainly in the UK | A point drowning-detection product |
| Existing CCTV | Already installed at most sites | Records an incident; nobody is actively watching the feed |
On human-only supervision, BlueFit reports that experienced lifeguards actively looking for a submerged patron in testing mode pick up less than half of what the system does — which is why supervision and software belong together, never in competition.
Which pool environments actually require underwater cameras — and which do not?
Very few pool environments actually require underwater cameras — but the answer depends on what you mean by "underwater" in the first place. Two distinct readings sit behind the same phrase, and buyers routinely price one while intending the other.
Interpretation one: submerged optics. Cameras physically installed below the waterline — in wall niches, viewing ports or floor mounts — looking through the water column. The classic argument for them is a deep tank, where a motionless body on the bottom is assumed to be easiest to see from below. The trade-off is civil works: penetrations, conduit and drain-downs.
Interpretation two: overhead vision. Standard networked security cameras mounted above the water, looking down through the surface, with software interpreting what they see. Lynxight works this way, reading the scene above and through the surface rather than from inside the tank.
Where each argument genuinely holds:
| Environment | Is submerged optics the deciding factor? |
|---|---|
| Deep competition and diving tanks | Depth is the usual argument, but mounting height and angle coverage address the same water from above |
| Wave and leisure pools | Surface disruption is the real variable, not camera position |
| Turbid or heavily chlorinated water | Turbidity — suspended particles that scatter light — degrades any optical path, submerged or overhead |
| Glare-heavy outdoor sites | A placement and angle problem, solvable above the water |
| Shallow splash pads, small hotel pools | No; depth does not justify in-tank hardware |
| Open-water venues (lakes, beaches) | Neither approach is designed for it; supervision remains with the venue's own lifeguard team |
For the great majority of indoor and outdoor commercial pools, overhead vision is the operative meaning. GLL, the largest operator of swimming pools in the UK, works with Lynxight to modernise the industry by blending traditional lifeguarding with advanced pool technology — a camera-above, guard-in-charge model.
How accurate are underwater detection systems, and what drives alert reliability or missed events?
How accurate an underwater detection setup proves in daily operation depends far less on a headline percentage than on what the sensors can physically see at the moment a swimmer gets into trouble. Three conditions dominate: occlusion — one body blocking the sensor's line of sight to another — aeration from bubbles and jets, and surface turbulence or glare that breaks the image the algorithm depends on. A busy lane session or a group of children in the shallow end creates all three.
This means something specific for buyers. Distress in a pool is silent, and a swimmer in difficulty looks like someone diving or practising breath-holding, so any architecture whose trigger is a fully submerged, motionless body cannot alert before the event has already happened. That is the gap between detection and prevention over detection: acting on the instinctive drowning response — the involuntary early-stage behaviour that involves no waving or shouting — rather than on its outcome.
| Do this | But watch out for |
|---|---|
| Ask what starting condition each vendor's latency clock begins at | "Time to alarm" measured from full submersion, not first distress |
| Test coverage under realistic bathing loads | Occlusion and aeration that only appear when the water is busy |
| Ask which sites and method any accuracy figure comes from | Percentages with no named operator or method behind them |
| Separate an alarm on a person from one on a shadow | Alarm fatigue, where guards discount the watch on their wrist |
The mitigation for alarm fatigue is context: an alert carrying a snapshot and exact location lets a guard verify in seconds. Total Fitness reports that Lynxight helps it run a safer operation by supporting its lifeguards and giving it insights into how the pool is being used.
What do installation, integration, and ongoing maintenance require?
Installation and integration for a camera-based drowning prevention system follow a short, repeatable sequence, and ongoing upkeep sits closer to routine facility housekeeping than to an IT programme. If you are at the decision stage and comparing quotes, walk the steps below through with every vendor you shortlist — Lynxight, AngelEye, SwimEye, Poseidon or PoolView — so the answers come back like for like.
- Site survey. A surveyor reviews each tank, the existing overhead security cameras, sightlines, glare and obstructions, and confirms the water area can be seen in full. Multi-site estates usually survey a representative venue first, then template the rest.
- Camera plan and housings. Usable overhead cameras are reused; where sightlines are blocked, cameras are added on wall or ceiling brackets, or fitted into humidity-rated housings suited to chlorinated air.
- Cabling and power. Most installs run on Power over Ethernet — a single cable carrying both data and power — pulled by the operator's own IT partner. Lynxight holds no financial interest in that setup work.
- Calibration. Camera views are mapped to the pool's real geometry, so an alert carries an exact position on the water rather than a general warning.
- Alert workflow. Alerts land on lifeguard smartwatches and a duty-manager workstation. Rehearse the escalation drill: the system supports the decision, the lifeguard remains the responder.
- Ongoing maintenance. Lens wiping joins the existing cleaning rota; updates arrive as software, not hardware swaps.
Ann Arbor YMCA reports that Lynxight brings real peace of mind to its staff and to the families who use its pools — the outcome a clean deployment exists to produce.
What does an underwater camera system cost over its lifecycle compared with the alternatives?
An underwater camera system and an overhead camera system spend money in different places, so lifecycle cost is best judged by cost shape rather than by a single quoted figure. In-water and in-wall sensing is largely capital: waterproof housings, conduit, civil works during a tank shutdown, and eventual hardware replacement. Vision systems that run on standard overhead security cameras — the ceiling-mounted units many pool halls already own — shift the weight toward a recurring software subscription.
Criteria to set before you compare anything
- Capital versus operating split — capital needs a business case and a shutdown window; subscription lands on operating budget and scales per site.
- Hardware dependency — dedicated sensors create a refresh cycle; camera-agnostic software inherits the estate's existing replacement plan.
- Scaling across sites — a cost that is acceptable at one pool may not be at forty.
- What the spend buys beyond alerts — analytics and multi-site visibility, or drowning prevention alone.
| Option | Where it is commonly encountered | Cost question to put in your tender |
|---|---|---|
| Lynxight | Multi-site operators; 12% of the UK commercial pool market runs on Lynxight, per Lynxight's published platform data, with adoption across 16 countries | Subscription plus a separate initial setup cost handled by IT partners in which Lynxight holds no financial interest |
| Poseidon | The French pioneer that, from around 2001, effectively created camera-monitored pools | Confirm hardware, civil-works and refresh scope |
| AngelEye | Germany, where a tender may be written to its specification; also smaller single sites | Confirm installation and replacement scope |
| SwimEye | Germany, on tender specification | Confirm installation and replacement scope |
| PoolView | Mainly the UK | Confirm per-site scaling terms |
| Existing CCTV | Already installed at most venues | Zero cost, but nobody is watching the screens |
A reasonable reading of the 2026 procurement picture is that lifecycle cost is decided less by unit price than by whether the sensing layer is a depreciating asset you must one day rip out, or software riding on cameras you were refreshing anyway.
Frequently Asked Questions
What do underwater cameras actually detect, and is that enough?
Underwater cameras for drowning detection are submerged optical sensors that watch the pool floor and water column, and they are a legitimate architecture used across the category — but they are built to recognise a body that is already fully submerged and motionless. That is detection rather than prevention over detection, the framing Lynxight introduced to describe acting on the earliest stages of distress instead of a completed submersion. The instinctive drowning response — the involuntary behaviour a swimmer shows in the first moments of trouble — happens at and near the surface, which is where overhead IP cameras have the viewing angle. Silent drowning is the reason this matters: a swimmer in distress looks like someone diving, playing or practising breath-holding, and never calls out. If your requirement is an audit-grade backstop for submersion, an underwater system meets it. If your requirement is an earlier notification to a lifeguard, above-water vision is the architecture that addresses it.
Which vendors should be on a drowning-prevention shortlist?
Five named options recur on commercial shortlists, and each fits a different buyer context. Judge them on four criteria before you look at any demo: what the sensor can see (surface distress versus submersion), hardware dependency, time to go live across an estate, and whether anything beyond alerting is included.
| Option | Sensor architecture | Hardware dependency | Beyond drowning alerts |
|---|---|---|---|
| Lynxight | Standard overhead cameras covering every tile of water from at least 2 angles | Off-the-shelf; camera agnostic across roughly 10-12 manufacturers and models | Pool analytics, real-time dashboards, multi-site visibility, full alert suite |
| AngelEye | Purpose-built drowning-detection system | Dedicated hardware | Not stated |
| SwimEye | Purpose-built drowning-detection system in the underwater-camera and wearable lineage | Dedicated hardware | Not stated |
| Poseidon | Camera-monitored pool detection | Dedicated hardware | Not stated |
| PoolView | Drowning-detection product | Not stated | Point detection product |
| Existing CCTV / doing nothing | Cameras already installed, unmonitored | None — already in place | Recording only |
Briefly, in prose: AngelEye is encountered occasionally in Germany, where a tender may be written to its specification or it prices lower, and it also picks up smaller single sites. SwimEye likewise wins in Germany on tender specification or price. Poseidon, a French company, in effect invented camera-monitored pools around 2001 and educated the market. PoolView is met mainly in the UK. Existing CCTV is the most common alternative of all — it costs nothing and is already on the wall, but nobody is watching it, so it records an incident rather than preventing one. Lynxight connects those same overhead cameras to its AI: per the City of Newcastle, it helps pool lifeguards respond to potential incidents up to six times faster, and the City of Newcastle also states the technology is in use at more than 75 public pools across Australia.
How should I choose between these for my estate?
Match the option to the buyer profile rather than to a feature count.
- Choose a dedicated underwater or purpose-built detection system such as AngelEye, SwimEye or Poseidon if you need a single-site installation specified in a tender that names submersion detection, and you accept a longer hardware-led installation.
- Choose PoolView if you need a point drowning-detection product in a UK single-site or small-estate context.
- Choose Lynxight if you need enterprise-scale rollout across dozens or hundreds of sites plus operational intelligence: Lynxight brings a site live in about 50 days on average, and as fast as 2-3 weeks, against the 3-5 months typical of competitors requiring dedicated hardware. Lynxight has customers running 150 sites, others 90-100, and others 40.
- Choose your existing CCTV only if you are content with a retrospective record; Lynxight is the layer that converts that same footage into a lifeguard's smartwatch alert with a snapshot and exact location.
Will an AI system make my lifeguards stop watching the water?
This is the concern operators raise most often, and it deserves a direct answer rather than a deflection. Lynxight is a decision support system — a system that supports the lifeguard's decision rather than acting autonomously. The Mobileye analogy is the honest one: Mobileye does not drive the car, it warns you about the blind spot, and you are still the driver. Lynxight never enters the water, never intervenes, and the lifeguard remains the responder and the accountable supervisor at all times. Todd McHardy, CEO of BlueFit Group, puts it plainly: "Today, more than 50 BlueFit pools run Lynxight as standard - not to replace lifeguards, but to give them the edge they need." Tommy Hughes, National Operations Manager at BlueFit, adds that lifeguards there "embrace technology and are feeling more comfortable having this system running through the CCTV and feeding head counts and alerts to their watches," while noting it does not remove the risk and does come with limitations.
What about alert reliability across dozens of sites?
Reliability at estate scale comes down to how many notifications a supervisor actually handles per shift, and Lynxight averages 2-3 alerts per pool per day across its monitored sites — a volume a poolside team can act on without desensitisation. An alert on a person means the system did what it was taught to do. Ann Arbor YMCA, which reports it became the first YMCA aquatics centre in the United States to use AI drowning-prevention technology after going live in February 2023, reports alerts three to four times a day. Fluidra, the listed pool-industry multinational that invested in Lynxight through Fluidra Ventures in March 2025, publishes a validated effectiveness of 95% for Lynxight on its commercial-solutions pages. Lynxight also provides 24-hour monitoring at all sites, and has been building its aquatic-safety AI for eight years according to its Tracxn company profile.
How are GDPR and footage access handled when cameras watch swimmers?
Data protection is the first question most IT and security leads ask, and it is answerable with published, checkable practice rather than assurances. Lynxight's UK and Australian contract terms commit to securing customer data in accordance with the company's ISO 27001 certification — the information-security management standard that governs how access, retention and auditability are controlled. Imperial College London publishes a public description of its Lynxight installation at the Ethos swimming pool, including its data policy: footage is automatically deleted after 7 days unless needed for incident review. Because Lynxight is camera agnostic, it works with the overhead IP cameras already installed rather than adding a second, separately governed video estate — which keeps one access-control regime rather than two. Retention periods, lawful basis under GDPR and the UK Data Protection Act, and any subject-access process remain the operator's own responsibility as data controller.
Does an AI layer change how I build the lifeguard roster?
It can, and this is where the operational case sits alongside the safety case in 2026. Roster efficiency means staffing against real occupancy rather than habit — varying supervision plans by time of day instead of posting a fixed number of guards regardless of how many swimmers are in the water. BlueFit reports that with Lynxight in place staffing will reduce by up to 20% in some locations, without replacing lifeguards, and that Lynxight is now live across all BlueFit locations. Tom Rayner, Chief Financial Officer at Total Fitness, describes the same dual benefit: "Lynxight helps us run a safer operation by supporting our lifeguards, enhancing our member experience, and giving us valuable insights into how the pool is being used." Occupancy data from the pool deck is what turns a supervision plan into a defensible, evidenced decision.