A Real-Time, Vision-Based System for Badminton Smash Speed Estimation on Mobile Devices
This work addresses the need for affordable performance analytics in badminton, particularly for amateur and recreational players, though it is incremental as it applies existing methods like YOLOv5 and Kalman filters to a new domain.
The paper tackled the problem of expensive and inaccessible performance metrics in badminton by developing a real-time, vision-based system using smartphones to estimate smash speed, achieving a cost-effective and user-friendly solution for players.
Performance metrics in sports, such as shot speed and angle, provide crucial feedback for athlete development. However, the technology to capture these metrics has historically been expensive, complex, and largely inaccessible to amateur and recreational players. This paper addresses this gap in the context of badminton, one of the world's most popular sports, by introducing a novel, cost-effective, and user-friendly system for measuring smash speed using ubiquitous smartphone technology. Our approach leverages a custom-trained YOLOv5 model for shuttlecock detection, combined with a Kalman filter for robust trajectory tracking. By implementing a video-based kinematic speed estimation method with spatiotemporal scaling, the system automatically calculates the shuttlecock's velocity from a standard video recording. The entire process is packaged into an intuitive mobile application, democratizing access to high-level performance analytics and empowering players at all levels to analyze and improve their game.