Determining NHS Room Occupancy Levels Through The Use of IoT Devices and Computer Vision

Pattern recognition identifying individuals within a room

The Rotherham NHS Foundation Trust (TRFT) has expanded its valuable and successful collaboration with The Curve. Their goal is to explore new and innovative ways to manage data more effectively and ensure real-time data is at the forefront of its decision making process.

Building on the effective design and implementation of the “Escalate” system, TRFT challenged The Curve to investigate possible solutions for determining the number of individuals in a room without human intervention, while maintaining an affordable hardware cost. The objective was to accurately and securely calculate the occupancy of rooms within the Trust, enabling improved and more accurate utilisation.

This led to the start of the “Room Occupancy Project” – the premise was simple: record the number of occupants within a room at the Trust without the need to manually count the number of people within each room. This was a key factor as results often consisted of subjective opinions such as; “the room was 90% full” rather than factual information. This led to inaccurate data which did not represent the actual environment and room occupancy levels, negatively impacting staff decisions.

Our approach involved using a series of IoT devices and systems to create a smart way in which data could be gathered remotely. This consisted of utilising a low-cost Single Board Computer (SBC) device fitted with a wide-angle camera lens which could be connected to a network and instructed to take photos of the room in question at intermittent periods throughout the day.

Images of the room are processed using an advanced AI computer vision model to determine how many people are present. The resulting data is captured and stored centrally, while the image itself is discarded for confidentiality purposes. This data is then used to visualise how the rooms are being used over time using a combination of tables and graphs.

A single board computer (SBC)

The platform has been developed using .NET Core 6, backed by an SQL Server database and using React.js for the front-end webviews. Azure Entra ID (formerly Active Directory) is used for authentication, and Azure Function services for scalable image processing.

Embracing these low-cost IoT solutions can help Trusts and other NHS organisations record more data and information than ever before. With the ability to collect more relevant and accurate data, this can lead to a more informed decision-making process that not only benefits the day-to-day operational management of a Trust but also the patients they care for.

Technologies used within this project:

.NET 6.0 SDK (LTS)

NodeJS 18 (LTS)

ReactJS 18.2.0