The 4th gold medal of the Chinese team is coming!
In the men's 500-meter speed skating competition of the Beijing Winter Olympics on the 12th, Gao Tingyu broke the Olympic record with a time of 34.32 seconds and won the first winter Olympic gold medal in Chinese men's speed skating, which was also the fourth gold medal of the Chinese sports delegation at the Winter Olympics.
Gao Tingyu in the game. Photo by China News Service reporter Li Jun
Behind this precious gold medal is the unremitting efforts of athletes and coaches, as well as years of scientific training. In sports, improving the accuracy of elements such as power and posture is the key to improving the efficiency of training, and it plays an important role in sports performance. But how to guide athletes to train precisely?
Traditional competitive training has long been constrained by a lack of data, and coaches usually coach athletes through naked-eye observation and personal experience. Lack of data means lack of models to guide precision training.
With the development of video acquisition and analysis technology, it has become possible to train athletes through sports posture data processing and sports performance analysis.
Intel 3DAT 3D Athlete Tracking Technology is a tool that can accurately and analyze athlete biomechanical data in real time to help athletes train efficiently. This technique has also been applied to training by the Chinese speed skating team.
3DAT uses the camera to shoot the athlete's training process, and the athlete does not have to wear sensors and other equipment that affects the motion effect, the system sends the video data to the cloud, analyzes it on the third-generation Intel Xeon Scalable processor with built-in Intel deep learning acceleration technology, and directly through the artificial intelligence computer vision algorithm, extracts the athlete's key bone point information, the posture of key parts of the body, the movement trajectory and position information from the video source, and reconstructs it in three dimensions. Biomechanical data is generated while the 3D model is reconstructed, and accurate motion performance analysis is output. These data and analysis can help coaches develop scientific and correct training methods and provide practical and useful technical support for athletes' daily training.
In the training process of speed skaters, take the starting and straight acceleration stages as an example. 3DAT will extract the key bone point information, key part posture, motion trajectory, etc. of speed skaters from the training video source for three-dimensional reconstruction, and then analyze the data through artificial intelligence to generate a parameter set of athlete biomechanical data. These parameter sets are an important basis for evaluating athletes and guiding adjustments.
For example, in the starting stage, the 3DAT technology can generate a set of parameters such as the athlete's cadence, ice pedaling time, suspension and ground time, knee angular velocity, starting angular velocity, etc., and the analysis of these data can understand the impact of starting attitude on speed.
In the straight phase, the rotation angle and angular velocity of the core joint, knee joint, and hip joint can be obtained through 3DAT, as well as the trajectory and acceleration of the athlete's centroid, which can help the athlete analyze the impact of posture on the quality and efficiency of the ice pedaling.
The accumulation of these data is essential for sports research and athlete training. Technicians can analyze the data of some outstanding athletes and generate a batch of data models, and these excellent data models will become the reference targets and comparison objects for athletes and coaches.
The blessing of these "black technologies" allows athletes to give full play to their potential in training, and also injects the power of science and technology into the Olympic Games. (End)