Teaching and evaluation methods for track and field long jump based on intelligent assessment technology

In view of the shortcomings of the existing intelligent sports teaching technologies in terms of the accuracy of action assessment for continuous movements and the effectiveness of teaching improvement, a standard long-jump teaching and assessment model has been proposed to conduct long-jump action assessment and teaching. Specifically, deep learning algorithms including YOLOv5, DeepSORT, the OpenPose framework, Bidirectional Long Short-Term Memory, and the Dynamic Time Warping algorithm are employed to measure run-up speed and evaluate long jump movements across the take-off, flight, and landing phases. In addition, a multi-level architecture comprising an intelligent estimation layer, an interactive instruction layer, and a teaching feedback layer is constructed to support adaptive physical education. Experimental results show that the intelligent estimation module achieves an evaluation accuracy of 90.50% for the take-off movement, demonstrating high detection precision. Therefore, during the eight-week student training period, the scores in the Take off stage increased by 9.33 points, the scores in the Leave the air stage rose to 10.36 points, and the scores in the Landing stage reached 7.34 points. Comparatively, these results are better than those from other comparison models. The efficiency of the proposed model is proved to enhance movement evaluation accuracy and teaching effectiveness.