投篮动作矫正系统
Shooting Motion Correction System
DOI: 10.12677/sea.2026.153035, PDF,   
作者: 张锡尧：中南大学自动化学院，湖南 长沙
关键词: 计算机视觉；多目标跟踪；姿态估计 ；Computer Vision； Multiple Object Tracking； Human Pose Estimation
摘要: 针对篮球爱好者缺乏便捷专业投篮动作矫正工具、传统人工矫正效率低且精准度不足的问题，本文研究基于计算机视觉的投篮动作矫正系统。该系统以普通设备拍摄视频为输入，集成多目标跟踪(MOT)、2D人体姿态估计(HRNet)等技术构建人体运动分析流水线，笔者主要负责优化多目标跟踪、设计投篮人筛选算法、开发ShotAnalyzer类有限状态机，及基于HRNet提取17个人体核心关节点并计算相关生物力学指标，深度运用相关专业知识完成研究。研究结果表明，系统成功实现投篮人精准锁定、多维度生物力学数据清晰输出及出手瞬间精准识别，运行稳定且能有效处理普通设备视频数据。综上，该系统解决了传统投篮动作矫正的短板，为系统后续完善提供技术支撑，可为篮球爱好者提供便捷专业的矫正参考，具有实用价值。
Abstract: To address the problem that basketball enthusiasts lack convenient and professional tools for shooting motion correction, and traditional manual correction is inefficient and inaccurate, this paper studies a shooting motion correction system based on computer vision. Taking videos captured by ordinary devices as input, the system integrates Multiple Object Tracking (MOT), 2D human pose estimation (HRNet) and other technologies to construct a complete human motion analysis pipeline. The author is mainly responsible for optimizing multi-object tracking, designing a shooter screening algorithm, developing a finite state machine of the ShotAnalyzer class, extracting 17 core human joint points based on HRNet and calculating relevant biomechanical indicators, and completing the research by deeply applying relevant professional knowledge. The research results show that the system has successfully achieved accurate positioning of shooters, clear output of multi-dimensional biomechanical data and precise identification of the ball release moment, running stably and effectively processing video data from ordinary devices. In summary, the system solves the shortcomings of traditional shooting motion correction, provides technical support for the subsequent improvement of the system, can provide convenient and professional correction references for basketball enthusiasts, and has practical value.
文章引用：张锡尧. 投篮动作矫正系统[J]. 软件工程与应用, 2026, 15(3): 373-383. https://doi.org/10.12677/sea.2026.153035

参考文献

[1] Cabarkapa, D., Fry, A.C. and Carlson, K.M. (2022) Review of the Basketball Jump Shot. Journal of Human Kinetics, 81, 147-158.
[2] Miller, S. and Bartlett, R. (1996) The Relationship between Basketball Shooting Kinematics, Distance and Playing Position. Journal of Sports Sciences, 14, 243-253. [Google Scholar] [CrossRef] [PubMed]
[3] Bewley, A., Ge, Z., Ott, L., Ramos, F. and Upcroft, B. (2016) Simple Online and Realtime Tracking. 2016 IEEE International Conference on Image Processing (ICIP), Phoenix, 25-28 September 2016, 3464-3468. [Google Scholar] [CrossRef] 
[4] Wojke, N., Bewley, A. and Paulus, D. (2017) Simple Online and Realtime Tracking with a Deep Association Metric. 2017 IEEE International Conference on Image Processing (ICIP), Beijing, 17-20 September 2017, 3645-3649. [Google Scholar] [CrossRef] 
[5] PaddlePaddle (2023) PaddleDetection: Object Detection and Tracking Toolkit Based on PaddlePaddle.
https://github.com/PaddlePaddle/PaddleDetection
[6] Cao, Z., Hidalgo, G., Simon, T., Wei, S. and Sheikh, Y. (2021) Openpose: Realtime Multi-Person 2D Pose Estimation Using Part Affinity Fields. IEEE Transactions on Pattern Analysis and Machine Intelligence, 43, 172-186. [Google Scholar] [CrossRef] [PubMed]
[7] Sun, K., Xiao, B., Liu, D. and Wang, J. (2019) Deep High-Resolution Representation Learning for Human Pose Estimation. 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, 15-20 June 2019, 5693-5703. [Google Scholar] [CrossRef] 
[8] Yan, S., Xiong, Y. and Lin, D. (2018) Spatial Temporal Graph Convolutional Networks for Skeleton-Based Action Recognition. Proceedings of the AAAI Conference on Artificial Intelligence, 32, 7444-7452. [Google Scholar] [CrossRef] 
[9] Shi, L., Zhang, Y., Cheng, J., et al. (2019) Two-Stream Adaptive Graph Convolutional Networks for Skeleton-Based Action Recognition. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Long Beach, 15-20 June 2019, 12026-12035.

为你推荐