Our swimming pool is fitted with Lynxight, an AI-assisted system to further support and enhance the safety management of our pool. This system will detect behaviours that may indicate someone requires assistance. By providing real-time alerts and analytics, the system will support lifeguards to improve your safety.

Please be assured that this upgrade is designed to work alongside our dedicated lifeguards, not replace them.

Key Information

  • Overhead cameras will point directly down at the pool. These cameras are designed to track movement only and use a colour-coding system to identify activity patterns.
  • The system cannot identify individuals, and data or footage will only be accessed if an incident occurs. To access the footage, we must submit a formal request with justification.
  • All footage will be automatically deleted after 7 days unless it is needed for incident review.
  • Before the system goes live, it will require a learning period to understand typical behaviour in the pool. The official launch date of the system depends on how long this takes. 

Thank you for your understanding and support as we implement this technology to improve your pool experience.

If you have any questions that are not covered on this page, please contact us via ASK.

Frequently Asked Questions

How does Lynxight AI work?

Spotting

This algorithm identifies, detects and differentiates between pool users and distinct types of objects, accounting for various situations, including:

  • Spotting objects above and below water.
  • Spotting moving and static objects.
  • Mitigation of water glaring impact, refractions, ripples and lighting conditions.

Tracking

This algorithm analyses movements and locks-in on every spotted pool user and object, simultaneously, from the moment of contact with the water until pool exit. It accounts for various situations, including:

  • Tracking objects above and below water.
  • Tracking under occlusions.
  • Tracking in high-density scenarios.
  • Collection of individual representations (locations, behaviour, camera angles and lighting).

Classifying

This algorithm determines pool user attributes and activity. Most notably, it serves to study pool user behaviour and actions in order to identify distress behaviour, accounting for various attributes, including:

  • Pool user risk assessment (based on water behaviour).
  • Pool user demographics.
  • Pool user activity.
  • Object interrelations.
  • Classification of pool inflatables and swim aids.

Deep Learning

Supported by a comprehensive dataset of annotated pool footage, a variety of deep learning algorithms are employed throughout the machine vision core data flow in different capacity, including:

  • Proprietary deep convolutional networks.
  • Unique data labelling and network training technology.
  • Computational optimisation for enhanced real-time performance.
Does the system track location and behaviour of everyone, everywhere?

The system tracks pool users within pool area only.

Does the system collect any kind of personal data?

The system does not collect any personal data of individuals (e.g. individual behaviour, face recognition, etc).

Does the system replace the need for lifeguards?

No, an automated monitoring and detection system does not replace the need for lifeguards. The main purpose of the system is to support and alert lifeguards to promptly identify and respond to a developing emergency. 

 

What are the benefits of having the system?

Lynxight is an extra layer of protection for pool users as it assists lifeguards with monitoring and supervising the pool area. This system enables lifeguards to react promptly to potential incidents or emergencies by providing timely alerts. The use of the system offers additional oversight for the lifeguards and ensures a higher level of safety and vigilance in the pool area.

Booking and Cancellation Information

Due to an an ongoing issue with users booking sessions and then not attending without cancelling beforehand, we have introduced a non-cancellation penalty. Any members who miss multiple sessions within a 7-day period will have their booking ability suspended for the following 7 days. Any pre-booked sessions for this timeframe will be cancelled and fully refunded.

Please note: all sessions can be cancelled up until they are due to start via MyLeisureHub. Pay-as-you go members can cancel via ASK (Imperial Staff and Students) or by emailing move@imperial.ac.uk (Other Members).