As the global economy is hit hard by the outbreak of COVID-19, it is everyone’s responsibility to help contain the crisis. Working from home, timely quarantining suspected individuals, as well as employing innovative technologies and cloud computing are some of the effective measures that help prevent the spread of the virus. The Data Science and Innovation Team of CITIC Telecom CPC thus had a thought of developing a body temperature measuring solution to address the increasing needs. With the regional teams’ effort, the solution is brought to reality now. Let us share this unique experience with you here.
Symptoms associated with COVID-19 include coughing, breathing difficulty, fever and shortness of breath. As a global ICT company with over 1,000 staffs, 26 offices across different regions, CITIC Telecom CPC attaches the highest importance to the safety and health of employees and customers. In light of the shortage of temperature monitoring devices since the outbreak and the crowded workplace environment, in addition to taking all protective measures, we also proactively addressed the problem by mobilising our innovation teams worldwide to collaborate and develop a thermal imaging solution for body temperature monitoring in just a few days.
Human body temperature monitoring systems available in the market are priced at USD5,000 to USD30,000, and they are in serious shortage currently. As such, our team started the process by purchasing industry-standard devices featuring both thermal imaging and general cameras. These competitively priced products can accurately detect the temperature of an object within 1 metre and allow easy adjustment of reading. With the tool kits developed by manufacturers and the remarkable capabilities of our Mainland, Hong Kong and Europe teams, CITIC Telecom CPC speedily developed a prototype of the body temperature monitoring system embedded with an alert function.
Leading the innovation team is Mr. Daniel Kwong, Chief Information and Innovation Officer of CITIC Telecom CPC. “Many places need to check the body temperature of every visitor with a handheld electronic thermometer, and it may increase the risk of cross-infection. However, a professional body temperature monitoring system is expensive and in short supply, and it takes much time to install," said Kwong.
The mechanism of thermal imaging technology is showing the temperature distribution of a remote object by converting its infrared radiation level to a temperature reading. But a human body temperature monitoring system must have a preset threshold value (i.e. a temperature level higher than the average level) so that a warning will be issued once a higher value is detected. However, merely setting a fixed threshold may cause errors or false detections.
Therefore, finding the right thermal imaging equipment was just the first step, which had to be followed by the development of a real-time detection and alert system. From sourcing the equipment to developing the system prototype, the team succeeded in building a body temperature measuring solution at a low cost within one week. They also augmented its performance by integrating AI to reduce detection errors, improving the solution for commercial use while ensuring the cost-effectiveness.
▲ Thermal imaging temperature monitoring system
Instead of scanning a single person, the thermal imaging technology aims to scan multiple targets at the same time. Computing power of the system is thus a challenge. The body temperature measuring solution developed by the team can simply run on a common personal computer, achieving cost-effectiveness without compromising the computing power requirement. It can receive and process the streaming signals from both the thermal imaging and general cameras simultaneously, and to issue a real-time alert if necessary. Since the general Windows operating system is not a real-time system, the team overcame the challenge by developing an application of real-time execution through multithreaded data processing that could maintain the speed.
"CITIC Telecom CPC has been investing in AI visual computing and machine learning to facilitate automated operation and improve the user experience in recent years. Our face recognition solution based on visual computing has already reached the commercial level and is ready for enterprises use. Further, our machine learning solution can help our customers to predict the probability of their systems or devices failure, and we can even base on the findings to detect our customers server upgrade needs and recommend the best solution for them. In fact, we have been developing many sophisticated technologies which can be widely used for business development. The technologies are our important assets and now, under the COVID-19 outbreak situation, it’s time for us to make use of them to address the needs," noted Kwong.
Big Data and AI enable more potential
With a higher visual computing accuracy and the face recognition technology, the system can specifically target the forehead of the face to ensure a more accurate measurement. This can avoid false alert, say when a visitor is holding a hot drink and pass-by the camera.
▲ Daniel Kwong, Chief Information and Innovation Officer, showcases the accuracy of the monitoring system.
As the average human body temperature has been declining since the Industrial Revolution, the definition of “normal body temperature” has also changed. Besides, our normal body temperature may fluctuate as our body regulates it according to the internal and external environments. Therefore, the alert value should not be an absolute value. It is essential to identify an abnormal fluctuation range so as to find out more meaningful alert values.
To achieve this, the innovation team is making use of cloud computing to empower big data analysis through machine learning. It is expected that, through continuous learning, the abnormal temperature range can be determined precisely and the system accuracy can attain a new height.
▲ When an abnormal temperature is detected, the system will flash red and activate an alarm. This photo shows the detection of the flame temperature of a lighter.
Another benefit of adopting cloud computing is that it can work with smart devices, such as smartphones, seamlessly. If the system detects a person with abnormal body temperature, it will take pictures of the person and notify the responsible officer through cloud computing and mobile network immediately.
Privacy is also our top concern. Notifications are placed in our offices to inform our visitors that their body temperature data will be collected for safety purpose. Meanwhile, the system also protects privacy technically as it can be programmed to target a specific visitor with abnormal body temperature and share with the responsible officer the relevant video captured within one second before and after the visitor passing in front of the camera for further review. It minimises the chance of capturing irrelevant visitors’ information (i.e. those who are with normal body temperature).
CITIC Telecom CPC is committed to offering world-class ICT solutions to companies of all sizes. The development of the latest body temperature measuring solution was not driven by profit but our hope to help combat COVID-19 with technologies and help businesses resume operation as soon as possible. Since we have received many customer enquiries so far and realised that many of them are faced with the same challenge, we decided to share our experience and insight.
Let’s stand for each other in adversity and overcome the crisis together!
Täname Teid päringu eest.