Apollon Mark: Ball Mark Visualization System with Preceding Mirror Control

Summary

Referee assistance systems are being used to judge the in/out of the ball in order to ensure fairness in ball sports. In tennis, in particular, many official tournaments have introduced systems that use high-speed multiple cameras to measure the three-dimensional trajectory of the ball and make in/out decisions based on the positional relationship of the ball three-dimensional trajectory with the tennis court lines. However, the bounce position of the ball estimated from the three-dimensional trajectory may have an error in the order of several millimeters from the actual ball mark. In fact, such multiple camera assistant tends not to be installed on clay courts, where ball marks are more likely to remain.

In this study, we propose a ball mark visualization system that uses a high-resolution camera to capture the minute physical deformation that occurs when the ball contacts the ground surface. We propose a ball mark visualization system based on high-speed prediction of the ball bounce position and preceding mirror control (Fig. 1). The system consists of the following three technical elements (Fig. 2). (1) Measure the three-dimensional trajectory of the ball using multiple high-speed cameras, estimate the motion information using a Kalman filter, and immediately predict the bounce position. (2) Control the line-of-sight direction of the high-resolution camera toward the predicted bounce position at high speed. (3) Take high-resolution texture images of the ground surface immediately before and after the bounce, and visualize the ball mark by the difference of the images. The system is named "Apollon Mark" after Apollo, the god of light and prophecy, because it uses prediction and light control technology to visualize the bounce mark.

This technology is unique in that it actively acquires a high-resolution image of the ground surface immediately before the ball bounces by utilizing high-speed motion prediction and optical control for ground surfaces such as grass where ball marks are difficult to recognize. In other words, the ball mark can be easily visualized by simply comparing an image of the ground surface immediately before the bounce with that of immediately after the bounce, which has sufficient resolution (Fig. 3). This system is expected not only to contribute to the development of the sports field as a fair and quick referee assistance system, but also to be applied to infrastructure applications such as the inspection of contact surfaces between moving objects and walls or the ground in the future.

Fig. 1 Concept.	Fig. 2 System configuration.
Fig. 3 Visualization of the ball bounce mark.

Reference

1. Himari Tochioka, Tomohiro Sueishi, and Masatoshi Ishikawa: Bounce Mark Visualization System for Ball Sports Judgement Using High-Speed Drop Location Prediction and Preceding Mirror Control, SICE Annual Conference 2023 (SICE2023) (Tsu, 2023.9.8)/Proceedings, pp.784-789 SICE Annual Conference International Award