Pitch Location Measurement Using a Single Camera and Radar Gun

This paper presents a practical and cost-effective method for measuring the pitch location in baseball using only a single camera and a radar gun. Pitch location, defined as the position at which a baseball crosses home plate, is a critical metric for both pitchers and batters aiming to improve performance. While commercial systems such as Trackman, Hawk-Eye, and Rapsodo can measure such data, their high cost and complexity limit accessibility, especially for amateur teams. To address this issue, the proposed method estimates the distance between the ball and camera by leveraging the ball’s physical displacement, as measured by a radar gun, and its corresponding pixel displacement in consecutive video frames. Additionally, the angular position of the ball is computed based on its location in the image plane. These parameters are then combined to reconstruct the 2D pitch location. The method is robust against rolling shutter distortion, with detection performance enhanced through a custom-trained YOLO model. Experimental validation was conducted on actual baseball fields, comparing results with those from Rapsodo and an existing monocular method. The proposed method demonstrated superior accuracy, with statistically significant reductions in estimation error. Given its simplicity, low cost, and field-tested accuracy, this approach offers a highly accessible solution for pitch location analysis, particularly in training environments lacking access to high-end tracking systems.