Real-time Localization of a Soccer Ball from a Single Camera
This provides a practical and accessible method for accurate 3D ball tracking in professional football environments, though it is incremental as it builds on prior single-camera approaches.
The paper tackles real-time 3D soccer ball localization from a single broadcast camera by introducing a multi-mode state model to accelerate optimization, achieving centimeter-level accuracy and low latency on standard CPUs, with performance comparable to multi-camera systems on a proprietary dataset of Russian Premier League matches.
We propose a computationally efficient method for real-time three-dimensional football trajectory reconstruction from a single broadcast camera. In contrast to previous work, our approach introduces a multi-mode state model with $W$ discrete modes to significantly accelerate optimization while preserving centimeter-level accuracy -- even in cases of severe occlusion, motion blur, and complex backgrounds. The system operates on standard CPUs and achieves low latency suitable for live broadcast settings. Extensive evaluation on a proprietary dataset of 6K-resolution Russian Premier League matches demonstrates performance comparable to multi-camera systems, without the need for specialized or costly infrastructure. This work provides a practical method for accessible and accurate 3D ball tracking in professional football environments.