ISO 20288 vs ISO 20380: understanding drowning detection standards

ISO 20380:2017 — "Public swimming pools — Computer vision systems for the detection of drowning accidents in swimming pools — Safety requirements and test methods" — was published in November 2017 by ISO/TC 83, the technical committee for sports and recreational facilities and equipment. It applies to public swimming pools and explicitly excludes domestic pools and basins with a surface area below 150 m².

What the standard specifies is precise, and narrower than many operators assume. It defines detection as the recognition of a stationary solid mass, fully immersed and lying at the bottom of the pool basin. Its performance requirements follow from that definition: the alarm must be set off within 15 seconds of the mass becoming stationary on the bottom, that set-off time must be built in so staff cannot change it, and the system must achieve a detection rate of at least 80% under the lighting conditions defined in the test method.

The standard also mandates process, not just performance: a technical study before installation (basin geometry, lighting, covered and non-covered areas), daily and half-yearly testing, a manual for trained staff, and maintenance. Conformity must therefore be assessed for a defined system, installation and operating regime; it is not a portable product badge.

ISO/FDIS 20288 — "Public swimming pools — Systems recognizing involuntary submersion of humans in managed aquatic facilities" — is a Final Draft International Standard under development within the same committee, ISO/TC 83. As of mid-2026 the ISO catalogue lists it at the FDIS stage, the formal approval phase that follows the enquiry ballot: national standards bodies have already reviewed and voted on the draft, and the near-final text is now registered for formal approval ahead of publication.

According to the catalogue entry, the draft specifies minimum requirements and test methods for systems that recognize the involuntary submersion of a person in managed aquatic facilities, including indoor pools, outdoor pools and free-water pools. The decisive shift is the object of detection: not a body already lying on the bottom, but the involuntary submersion event itself — earlier in the trajectory of a drowning incident.

One caveat matters for procurement: an FDIS is not yet a published standard. At this stage the text is essentially final, but until publication no vendor can legitimately claim certification against ISO 20288. Treat it as a reliable signal of where the industry is heading — very close now to a finished rulebook, but not yet one in force.

A drowning incident unfolds as a sequence: distress at the surface, involuntary submersion, sinking, and finally a motionless body on the pool floor. ISO 20380 addresses the final stage — confirming that a person has come to rest at the bottom. The scope of ISO/FDIS 20288 addresses the earlier stages: recognizing that an involuntary submersion is happening while it is happening.

ISO 20380 itself is candid about this boundary. Its introduction notes that such systems cannot detect every accident — a person floating at or just below the surface, for example — and that saving a drowning person always requires human intervention. Early-warning submersion recognition exists to close precisely the gap the published standard acknowledges.

The two approaches are complementary layers rather than competitors: bottom detection as a last line of defence, early-warning technology to move the alert forward in time.

Drowning is fast and often silent, and the outcome depends on several factors, including how quickly a person is reached and brought back to the surface. An earlier alert can give the lifeguard more time to locate the swimmer, cross the deck and intervene.

A bottom-detection alarm within 15 seconds of a body lying motionless on the floor is a genuinely valuable safety net, but by definition it begins after the submersion has already ended in the worst position. An early-warning system aims to alert staff during the submersion, so intervention can start before the person reaches the bottom. For an operator, this difference shapes response protocols, staff positioning and — most importantly — what you can realistically expect the technology to do.

A conformity claim should identify the tested system, configuration, conditions and evidence. A test report covering the standard’s requirements can be meaningful technical evidence; independent certification exists only when a certification body has actually issued a certificate for a defined scope. Ask to see the report or certificate and the scope it covers.

Compliance does not mean drownings are prevented, and it does not justify reducing supervision. ISO 20380 states explicitly that computer vision systems are designed to complement lifeguards and trained staff, not to reduce supervision or staffing levels. Drowning prevention remains a human system — trained people, clear procedures, appropriate pool design — in which detection technology is a support layer.

Nor does a certificate transfer responsibility wholesale to the supplier. Risk assessments, national bathing regulations and the facility's own safety organisation continue to apply. A standard tells you how a system was tested; it does not run your poolside response.

When evaluating a drowning detection system, a handful of direct questions will separate substance from marketing:

A serious supplier answers these precisely, with test protocols and documentation. Vague accuracy claims without a defined test method deserve caution: performance figures only mean something relative to a specified test.

Operators procuring today can prepare for ISO 20288 without waiting for publication: follow the project's progress in the ISO catalogue, favour system architectures built around early recognition of submersion rather than only end-state confirmation, and write response protocols that distinguish an early alert from a late-stage alarm.

Aqutron Vision is designed to identify conditions that may indicate a drowning emergency and route an alert to designated staff. Depending on the facility, it can use a camera-only configuration or add spatial sensors based on patented 3D technologies. Safety-critical processing is designed to run on site, without requiring facial recognition or biometric identification. Coverage, alert routing, privacy measures and performance criteria must be assessed and tested for the installed configuration.

Whatever technology you choose, the principle both standards share is the one that should guide your pool safety strategy: detection systems support human supervision — they never replace it. The strongest facilities layer trained people, clear procedures and well-tested drowning detection, so that when the unexpected happens, the alert arrives as early as possible.