A Technology to Make the Calls
at Home Plate

How would you design a technology to perfectly measure whether a runner is safe or out at home plate? In other words, your technology must determine which event occurred first, the runner touching the plate or the fielder tagging the runner, simple, inexpensive designs are preferred. Use the Club Jameco Whiteboard to tell us in a paragraph or three how you would design the system.


A Virtual Umpire Requires More Complexity
Than Meets the Eye

Submitted by: James Fumato

 baseball-k-zone
To perfectly assess the play at the plate you need to know who and what is in the area at the time. The actual algorithms are simple: determine first, whether the outer surface of a volume defining the runner is first contacted by the volume defining home plate or the volume defining the defensive hand holding the ball and then, determine if the defensive player controlled the ball after tagging the runner out. The program should ensure the runner stays within a defined base path and that the lead runner is not passed by the next runner.

The tricky part is dynamically surface mapping a player over the course of several seconds within a volume around home plate. To accommodate all players making a maximum stretch to reach home, we would need to set this volume about 10 feet from the base path and home plate. This would be a 4000 cubic foot space (20' x 20' x 10'). If each volume element (voxel) is 1 cubic cm, we are tracking 114 million of them per frame. Since an umpire "sees" about 20 frames per second, we need to capture 40 frames a second (4.6 billion voxels per second). Each voxel describes what's inside (grass, defensive uniform, catcher's mask, flying dirt, umpire).

Several camera angles will be needed to hope to capture what's going on. Because tracking motion is usually based on estimating the center of markers, not a surface cloud, new programs following the millions of surface points per object would need to be developed and would need to run in much faster processors (available in about 3 years).

The routine is simple because it's an extension of 2D collision detection programs. Because the play occurs within a large volume, events outside the umpire's peripheral vision could affect the outcome. An accessory umpire who patrols the first base line and has a different perspective on the play might be all that's needed to get more perfect calls at the plate (and at first base).

James Furmato, PhD, is a New York Yankees fan who grew up in New Jersey. He currently lives in Delaware and works in Philadelphia as chief engineer in the Gait Study Center of the Temple University School of Podiatric Medicine. He is a Doctor of Podiatric Medicine with degrees in Chemical and Biomedical Engineering.