Updated: Jun 1, 2020
With organisations looking at returning staff to the workplace, there has been much made about the possibilities of thermal cameras reading individuals temperatures to a high degree of accuracy. Whilst this technology certainly assists in finding obvious cases of raised body temperature, can the results be truly trusted?
An office building has set up a camera to measure incoming staff's temperatures. One staff member arrives in an air-conditioned taxi, one on a bike, and one on the underground. If all three walk immediately into a foyer for a temperature reading you will undoubtedly discover a range of temperatures and one of the three would probably give off a raised temperature alarm.
This is obviously a hypothetical scenario, but what value are these solutions if they are not accurate enough to be reliable?
Some of these issues can be mitigated through calibration and best practice. Allowing participants' time to return to an ambient temperature through a staggered entry policy, for instance, would assist the scenario highlighted. But how can we minimise false readings?
I have recently been watching online videos of people filing (at incredible speeds) past a thermal temperature camera. With one of the participants holding a hot glass against his head to mimic a raised temperature, this of course for a sales video gets picked up immediately by the camera. But is the forehead a good place to measure? Our proposed scenario above would indicate not, there are many variables that can affect this temperature, exposure to the sun, or cold weather, air conditioning to name a few.
Products from Hikvision and Dahua provide an additional solution to help the accuracy of the thermal image, a device known as a blackbody. This device helps to calibrate the cameras giving them a reference temperature but can be as little as £1300.00 and improves the accuracy from +/- 0.5°C to 0.3°C.
One of UTL's solution partners Grekkom Technologies specialise in developing video analysis solutions for thermal cameras. They have 15 years of experience providing the full chain of value, from R&D to post-sales support in thermal solutions, including perimeter protection, the marine environment, and also the oil and gas sector.
Together with an Axis Q2901-e thermal camera, their solution provides on-camera (embedded) software meaning no further equipment is required. They do not measure the highest temperature point within the face like many products in the market currently do, they scan an analyse several areas of the face including the eye duct which gives the best indication of internal temperature from a facial image.
Herein lies a requirement; to accurately detect and guarantee that accuracy individuals must stand one at a time for approximately 1-2 secs to scan the correct positions on the face. As a result, there are recommendations as to where these cameras are best suited and scenarios where they cannot be accurately relied upon, and therefore should NOT be used, such as:-
Areas of large people flow.
Near temperature controlling devices (air conditioning, or heaters).
Shopping centres and supermarkets.
Sports, or concert events.
Recommended areas of use would be indoors, with controlled temperatures ranging from 18-24°C and controlled people flow which would include:-
Passenger control - airports, cruises, public transport entrances.
Access control - Office buildings employee entrances, industry, hospitals and clinics.
Using Grekkoms HTC (Human Temperature Control) embedded on the Axis camera allows the solution to be integrated with access control, making barrier solutions an ideal scenario for the cameras and software to be deployed.
So undoubtedly there are benefits to temperature control when used in the correct way.
None of these products can identify Covid-19, only a raised temperature which can indicate a fever.
For more information on this solution or any other please contact us.
0207 1234 101