(Fighting against New Coronary Pneumonia) Artificial Intelligence Temperature Measurement "Hard Technology" Meets "School Tide"
China News Service, Beijing, April 28th Question: Artificial intelligence temperature measurement "hard technology" meets the "school wave"
China News Service reporter Zhang Su
How to keep the “red line” of school security in place after school resumes one after another? Facing the need for temperature measurement on campus and kindergarten, the scientific research team born out of the Computer Department of Tsinghua University and the Institute of Automation of the Chinese Academy of Sciences accelerated the development of infrared thermal imaging temperature measurement robots.
It is not new to make temperature measuring instruments based on infrared thermal imaging technology, but the main feature of this temperature measuring robot is artificial intelligence. Min Haibo, the representative of the developer and the head of the Bingo Intelligent Team, said in an interview with the China News Service on the 28th that non-contact thermometers used in traditional scenes such as airports and train stations are expensive and difficult to cover more places, especially Cost-sensitive schools and kindergartens. For a long time, the school and the garden have been accustomed to arranging personnel to hold the thermometer at the door for testing, which is time-consuming and laborious, the temperature measurement accuracy is not ideal, and there is a risk of cross-infection.
Min Haibo and his colleagues have been dealing with educational institutions for many years, realizing this "pain point", and started to work on the "economical" non-contact thermometer in May last year, which was originally planned to be completed in one year. The sudden new coronary pneumonia epidemic has prompted them to speed up the implementation of technology. Recently, they have launched products and deployed them in some schools and kindergartens.
This temperature measurement robot uses low-cost hardware, but through deep learning, reinforcement learning to achieve algorithm compensation, to achieve better accuracy and stability. For example, Min Haibo said that the key component of adult fluid temperature detection equipment is the "black body". With it, the temperature measurement is stable and reliable, but the cost cannot be reduced. And due to the epidemic situation, the supply of "black body" production capacity is not available. Their approach is to mobilize 180,000 pieces of data, built-in artificial intelligence deep learning, thereby avoiding "black body" verification.
Min Haibo said that after field test and measurement agency measurement, the temperature measurement robot supports face recognition with masks in the interval of 1-2 meters, which can measure 3-5 people at the same time and 80-120 people per minute. The temperature error is positive Between minus 0.3 degrees Celsius. The temperature-measuring robot can also automatically track human faces and automatically alert people with abnormal body temperature. Comprehensive comparison of traditional non-contact thermometers, "half the cost" under the same application effect.
The R & D team will also integrate the temperature measuring robot into the educational service robot manufactured and sold in the form of a module, plus the functions of "attendance" and "video push", taking into account the education and management needs of the school and the garden. Zuo Lijun, director of the kindergarten in Shijingshan District, Beijing, said that the integrated solution of infrared temperature measurement and educational robot greatly improves the efficiency of morning inspection and attendance, and it is easy to use and easy to use.
During the epidemic, Min Haibo and colleagues also rushed to create an interactive intelligent teaching online teaching platform. "The education and training institutions that help cooperation enroll nearly 20,000 students." He said, "Cloud Classroom" is in the ascendant, but for many education and training institutions for children aged 2-8, it is very practical to transfer offline courses to online Difficult, some live broadcast classes are generally stiff, it is difficult to mobilize the enthusiasm of children. The main role of this platform is to improve the gamification, interaction and intelligence of the course on the one hand, "teachers can create interactive courses more easily by dragging and editing modules"; on the other hand, the background algorithm is based on the children ’s questions and emotions in the class , Active degree, concentration degree and other situations to generate reports, assist teachers to give evaluations and targeted suggestions, and then better implement teaching according to their aptitude.
"Our ultimate goal is to introduce artificial intelligence into the classroom, let robots take advantage of data collection and analysis, and truly assist preschool teachers." Min Haibo admitted that the two parts of artificial intelligence and preschool services seem to be simple to combine, and there are technical barriers that require innovation. Technology and cross-border thinking. Since 2015, the Bingo Intelligence Team has been working with the State Key Laboratory of Intelligent Technology and Systems of Tsinghua University to explore. So far, the total duration of the use of intelligent robots in kindergarten courses has exceeded 5 million hours.
They even set their sights on remote villages. Last year, the team used a teaching assistant robot to implement the “one village, one child” push-up course in Liangshan Yi Autonomous Prefecture, Sichuan Province, and demonstrated it in 320 local villages. Min Haibo explained that there is an extreme shortage of Chinese teachers in the local area, and some Yi children are not suitable for the class because of the low level of Chinese, and then choose to drop out. "Assisting Yi teachers to teach Mandarin with intelligent robots that will never leave their jobs, and can also track the situation of local Yi children for a long time." He said that this year he plans to bring this project to the Ganzi Tibetan Autonomous Prefecture and contribute to the fight against poverty in the field of education. . (Finish)