At present, the system has been applied to the movement technical analysis of speed skating, cross-country skiing, cold water jug, snowmobile and other projects, and has been popularized and applied to track and field throwing projects, long jump, badminton, weightlifting and other projects.
Xinhua News Agency, Beijing, January 8 (Reporter Wei Meng Jia) the reporter learned from Beijing Sports University that in preparation for the Beijing 2022 Winter Olympics, the school has undertaken and implemented the national key research and development plan of “science and technology Winter Olympics” “research on the characteristics of special abilities of winter athletes and key technologies of scientific selection. At present, the research of this project has broken through the key technologies that restrict the improvement of competitive performance ability of some winter athletes, especially the application of AI technology, which will make the training of athletes sports in our country more scientific and effective.
Zhu Xiaolan, director of science and technology department of Beijing Sport University, introduced that researchers have developed a “smart ice rink” system based on AI technology and can obtain motion information in real time. In the speed skating hall of Heilongjiang ice base, this system can automatically track the ice position of athletes without touching athletes, draw their tracks and calculate speed-related indexes, it is convenient for athletes and coaches to grasp the training effect, physical fitness distribution and competition skills and tactics in time. The system will be used normally during winter training this year and will be promoted to various speed skating venues.
It is reported that the “smart ice rink” system adopts AI machine vision algorithm and panoramic picture splicing technology to monitor various sports information of speed skaters, so as to master the instantaneous speed of athletes during sliding, acceleration, sliding track and other motion information.
Zhu Xiaolan said that researchers have tried a wide range of video shooting the image and splicing technologies and have successfully applied them to the ultra-complex environment of steel snowmobiles, it realizes the reproduction of three-dimensional movements in ultra-complex scenes and provides high-tech guarantee for high-level training of coaches and athletes.
At the same time, the research team also developed a large-scale three-dimensional motion technology analysis system based on AI technology. The system can not only accurately identify and calculate the three-dimensional coordinates of human Joint points and key points of instruments in motion, but also calculate a series of action technical indexes such as human Barycentric coordinate system, joint angle, speed and movement state of instruments, etc, it can meet the needs of action technical analysis of most ice and snow projects.
It is understood that this system uses AI automatic recognition technology to automatically recognize the posture of athletes, which can obtain 21 key points of human body without arranging any sensors on athletes, coordinate parameters of human motion link can be obtained, and three-dimensional movement data of athletes can be synthesized by using spatial calibration data.
At present, the system has been applied to the movement Technology Analysis of speed skating, cross-country skiing, cold water jug, snowmobile and other projects, and has been popularized and applied to track and field throwing projects, long jump, badminton, weightlifting and other projects.