How Thermal Cameras can Help Prevent the Spread of COVID19
Around the world, governments are responding to the unprecedented circumstances related to the coronavirus (COVID-19) epidemic. In many countries and regions, authorities have placed restrictions on their citizens movements and have increased guidance on the basic hygiene required to reduce the spread of the virus. The primary aim of this activity is to reduce the reproduction number (Ro ) of COVID-19 by limiting contact between groups of people as much as possible. Similarly, many government and healthcare authorities have provided guidance on the key symptoms associated with the disease. One of the key symptoms is an increased body temperature or fever. How can thermal cameras help? There are several activities and approaches being applied to help reduce the reproduction rate of COVID-19. These include self-isolation methods such as working from home, improved basic hygiene such as increased hand washing and the deployment of personal protective equipment (PPE) to reduce the prospect of infection. Similarly, when symptoms appear there is clear guidance on what to do next. Primarily this involves limiting social contact through self-isolation for up to 14 days. Medical professionals should be contacted digitally if symptoms persist or deteriorate. Ultimately, prior to any vaccine being available, the fight against COVID-19 is being led by the ability to detect symptoms and isolate people suspected of an infection. This is a combined effort between different key workers and technology applications. Thermal cameras can play a part in this coordinated approach. These cameras provide thermal imaging for body temperature solutions which can quickly and accurately identify people with elevated body temperatures, one of the key symptoms of COVID-19. These solutions can provide organizations with an additional layer of protection to their facility from increased exposure to the coronavirus. Organizations can then decide how best to deploy this information based on region, culture and the critical nature of the facility. In some circumstances a security officer may ask the person to scan their temperature using a medically approved sensor. In others, the person may be denied access to the facility. Ultimately, it is a decision for each organization on how best to deploy the solution. Thermal body temperature solutions An important distinction to make in the overall societal response to COVID-19 is that body temperature solutions are not a medical solution. They cannot identify the virus and they do not protect organizations or individuals from catching the virus. Thermal body temperature solutions are a tool that can support the identification of a key symptom of the disease. They can help organizations identify people showing these symptoms, but they do not diagnose or treat COVID-19. However, this does not mean that thermal body temperature solutions do not add value in the overall response. In fact, they provide a non-invasive method to check body temperature, can do this at faster rates than hand-held scanners and at a greater (potentially safer) distance. The deployment of these solutions in a facility may even encourage positive behaviour with staff more likely to stay at home when they are unwell with a fever. Thermal body temperature solutions require, at a minimum, a radiometric thermal camera to measure temperature differences in people entering the field of view. More advanced solutions will use blackbody devices to help calibrate the temperature measurement, especially in less controlled environments where the elements can influence the reading. AI (artificial intelligence) algorithms can also be integrated to help target the temperature reading on the most accurate part of the body, typically the forehead or near the eyes. The blackbody calibration tool consists of a target object whose temperature is precisely known and controlled. Specifically, this is important in human temperature measurement where accuracy to +/- 0.3 degrees Celsius is advised by many international standards organizations. By deploying the blackbody calibration tool, it is easier to establish an accurate relationship between gray level and temperature. Essentially there is known, fixed temperature object in the field of view which can be used to calibrate and measure all other objects’ temperatures. Using this method, false temperature alarms caused by environmental influence can be effectively reduced, and the accuracy of the thermal body temperature solution can be controlled to more precise parameters. However, monitoring accuracy does depend on the stability of the body temperature and it is recommended to install the system in a stable environmental condition to ensure that the skin temperature is stable. The emergence of AI technology, and specifically face detection algorithms, will play an important role in the evolution of these solutions too. Algorithms can help complete more accurate temperature tests. Cameras can do this by locating specific areas of the face such as the forehead or eyes, more accurately. This could be critical in the case of people wearing masks. Combining thermal cameras and facial detection can enable thermal body temperature solutions to combine accurate temperature scanning with the best face location to take the measurement from, improving the overall measurement accuracy. It should also be noted that the facial detection, as opposed to recognition, is used to improve the accuracy of the solution with better positioning of the measuring point on the face. It is not used to detect specific individuals and does not break privacy compliances (such as GDPR). While there remain challenges to the effectiveness of thermal imaging cameras for measuring human body temperature in public areas, especially when face masks are commonplace, the introduction of facial detection and AI can improve the accuracy of temperature scanning. Managing expectations for use Comparisons can be made between the current stage of the market for thermal body temperature solutions and another physical security technology – video analytics. Here, the expectation level for object detection or activity tracking algorithms was extremely high. The expectation was that video analytics would be near 100 percent accurate in spotting, identifying and tracking objects through the field of vision. However, analytics would sometimes misunderstand a scene, potentially alerting to the same object multiple times or mis-allocating an object – essentially false alerts. The reality was that these solutions…