AUTOMATED OR ASSISTED UMPIRING OF BASEBALL GAME USING COMPUTER VISION

Smart overview of the Invention

The patent application outlines methods and systems designed to automate or assist in the umpiring of baseball or softball games through the use of computer vision. The technology involves capturing video images of a batter and the ball using multiple cameras positioned at different angles, allowing for precise determination of the strike zone and tracking of the ball's trajectory. This system autonomously assesses whether a pitch is a "strike" or a "ball" based on the ball's interaction with the strike zone and the batter's swing attempt.

Technical Approach

The method involves utilizing video images from at least two cameras to track both the ball and the bat in real-time. Computer vision algorithms are applied to determine whether the ball intersects with the strike zone and whether the batter makes a genuine attempt to swing. The system can autonomously output whether a pitch results in a "strike" or a "ball," offering a more objective assessment than traditional human umpiring.

3D Space Tracking

The technology employs transformations to convert 2D video images into 3D space coordinates, allowing for accurate tracking of both the ball and bat. This enables autonomous determination of contact between bat and ball, as well as assessments of swings by comparing bat location against predefined planes aligned with home plate. The system can also assess changes in the ball's trajectory to determine contact events.

Determining Swings and Contact

The system can deduce whether a batter made a genuine swing attempt by evaluating intersections between bat location and a vertical plane at home plate, or by analyzing bat rotation angles. Additionally, it considers audio cues to verify contact, which can help identify foul tips. These determinations contribute to accurately calling pitches and understanding play outcomes.

Automation and Output

By leveraging multiple cameras and processors, this technology performs its functions autonomously, without human intervention. It outputs results via visual displays or audio speakers, providing real-time feedback on game events such as strikes, balls, fouls, and fair hits. This system aims to enhance decision-making accuracy in baseball games while reducing reliance on human judgment.