Bosch announces a human skin temperature detection solution that features built-in artificial intelligence (AI).
Data from the Human Skin Temperature Detection solution from Bosch provides actionable insights that can minimize illness spread by accurately identifying people with elevated skin temperature at checkpoints in office buildings, manufacturing plants and airports, according to the company.
The camera’s thermal module has a radiometric accuracy of +/- 0.5 °C / 0.9 °F. The solution is among the devices that officially comply with the stringent norms of the FDA’s April 2020 Emergency Use Authorization (EUA) Enforcement Policy for Telethermographic Systems guidelines and has been tested by an independent third-party lab for accuracy, stability and drift per International Electrotechnical Commission IEC80601-2-59:2017 test setup, according to the company.
The solution also provides a contactless measurement within a range of 2.5 to 4.5 meters (8 to 14 feet).
The Bosch Human Skin Temperature Detection solution is offered as a kit that contains a DINION IP thermal 9000 RM camera combined with a smart VIDEOJET decoder 7000 that also acts as a direct monitor output, a temperature reference device (blackbody), a special control keyboard and specific software that helps to detect anomalies in human skin temperature.
The camera’s built-in AI or Intelligent Video Analytics detects faces in the field of view, and the decoder’s embedded thermal algorithm will locate the hotspots on the face. For reliable screening, Bosch says individuals need to face the camera, remove glasses, hats, and other head and face coverings, and remain still. The solution will quickly detect multiple facial hotspots and indicate if an individual has an elevated skin temperature.
The system’s interface allows operators to access temperature alarms, temperature histograms, quick snapshot replay and alarm threshold configurations via a USB keyboard, without the need for an additional computer or software.
The solution also offers a special GDPR compliant public self-service mode, which will cover the person’s face and hide the screen’s absolute temperature data. By default, the decoder will automatically erase stored historical data after four hours to ensure the latest data is used for reference and sensitive data is not unnecessarily kept.
Depending on the application, operators can choose from three unique alarm profiles — absolute, relative or automatic alarm thresholds — to achieve better screening accuracy.
In a controlled environment, such as inside a building, operators can use the absolute alarm threshold. This profile measures the absolute skin temperature of the hottest spots on the face and triggers an alert when the preconfigured threshold is reached or exceeded.
When skin temperature may vary due to environmental conditions, like air temperature, operators can apply the relative alarm threshold. With this profile, the alarm temperature threshold setup is based on the average temperature of multiple persons plus x degree °C / F.
For applications that need to scan large numbers of people daily, operators can choose the automatic alarm threshold. This profile provides an alarm whenever the skin temperature is within a configurable fraction (alarm ratio) of the hottest previously observed values. The alarm threshold is automatically based on the rolling average of the population previously observed. All profiles offer either manual or automatic face detection.
Bosch says the solution’s built-in AI not only enables the detection of faces of which it measures the hottest spot. It also filters possible false triggers caused by heat sources, like a cup of coffee. The system continuously monitors all temperature data (both historical and current). If there are too many deviations, a new threshold is estimated so that an operator can update system settings accordingly.
This post premiered on our sister site, Security Sales & Integration.