本文提出了一种基于姿态识别技术的田径的短跑起跑的动作识别与评价系统。该系统通过使用RGB摄像头和深度相机获取运动员的运动数据，并利用深度学习算法进行姿态识别和动作识别。同时，系统还提供了基于运动数据的动作评价指标，用于评估运动员的技能水平。实验结果表明，该系统能够准确地识别起跑的准备姿势和预备姿势各个角度的大小，并提供有用的动作评价指标，为短跑运动员的训练和提高技能水平提供了更加直观的参考和反馈。

This paper presents a motion recognition and evaluation system for track and field based on posture recognition technology. The system obtains the movement data of the athletes by using the RGB camera and depth camera, and uses the deep learning algorithm for gesture recognition and motion recognition. At the same time, the system also provides action evaluation indicators based on sports data to evaluate the skill level of athletes. The experimental results show that the system can accurately identify the size of all angles of starting preparation and preparatory posture, and provide useful movement evaluation indicators, which provide more intuitive reference and feedback for sprinters' training and improving skill level.