With the rapid development of society and technology, the physical health of university students has become a critical concern, influencing both individual well-being and the national talent pool. This study employs an improved K-means algorithm integrated with machine learning models to analyze university students’ fitness data and develop personalized health intervention strategies. The enhanced K-means algorithm overcomes the limitations of traditional clustering approaches, leading to improved clustering accuracy and stability. Machine learning models—including Random Forest, decision trees, Gradient Boosting Trees, and logistic regression—were utilized to validate the clustering outcomes and to identify key health indicators associated with different student groups. Based on the clustering and model analysis, targeted intervention programs are proposed, such as strength training for groups with low muscular explosiveness, endurance training for those with stamina deficiencies, and flexibility exercises for groups exhibiting limited mobility. This integrated analytical framework provides a scientifically grounded tool for comprehensive health assessments and offers actionable, data-driven recommendations for student health management. Future research will focus on optimizing algorithmic performance, enhancing data diversity, and broadening the application scope to further improve the effectiveness and feasibility of health interventions.

The landmine press in a reliable and valid test for assessing upper body push strength. This study developed a markerless, non-contact vision-based system with an enhanced YOLOv8 model and mathematical modeling to measure four kinematic indicators during the concentric phase of the landmine press. By integrating a polarized self-attention mechanism, an improved C3k2 module, and an optimized SPPF structure, the system significantly enhanced detection accuracy and robustness for small targets at both ends of the barbell, achieving an mAP@0.5 of 0.995 on the test set. Agreement validation with the GymAware linear transducer across four loads (20–35 kg) in 247 trials showed strong correlations (r > 0.85) for peak velocity, mean velocity, peak power, and mean power. Although the vision-based method systematically overestimated velocity metrics due to differences in barbell length measurement, the bias was predictable. Moreover, it outperformed GymAware under high-load conditions (35 kg) by avoiding sensor drift and noise. The findings demonstrate that this vision-based system offers a reliable and practical solution for monitoring landmine press kinematics, suitable for both training and scientific research.

The landmine press is a reliable and valid test for assessing upper-body push strength. However, its application is constrained by the limitations of current mainstream monitoring technologies, such as linear position transducers (LPTs). These devices require physical attachment to the barbell, they rely on proprietary software, and their measurement accuracy can degrade under high-load conditions due to sensor drift and electromechanical noise. To address these limitations, this study developed a markerless, non-contact, and vision-based system using an enhanced YOLOv8-OBB model and a mathematical modeling framework to measure four kinematic indicators during the concentric phase of the landmine press. By integrating a polarized self-attention mechanism, an improved C3k2 module, and an optimized SPPF structure, the system significantly enhanced detection accuracy and robustness for the small targets at both ends of the barbell, achieving an mAP@0.5 of 0.995 on the test set. A method comparison study was conducted against a widely used LPT device (GymAware) across four loads (20–35 kg) in 247 trials. The results showed strong correlations (r > 0.85) for peak velocity, mean velocity, peak power, and mean power. Although the vision-based method systematically overestimated velocity metrics, the bias was predictable. Notably, under the highest load (35 kg), where LPT limitations are pronounced, the vision system demonstrated comparative stability, suggesting its potential advantage in mitigating sensor-related errors. The findings demonstrate that this vision-based system offers a reliable and practical alternative for monitoring landmine press kinematics, suitable for both training and scientific research.