Goal distance estimation in soccer game

Abstract

We examine the problem of object positioning in the 3D Euclidean space from uncalibrated images. It is well known that this can be accomplished if partial information about some points or planes in the observed scene are available. We consider the specific context of plotting the 3D position of a goal-bound soccer ball. In this context we know the position in the 3D Euclidean space of the straight lines defining the boundary of the goal-mouth structure. The method we consider handles pairs of uncalibrated images using the "plane + parallax" (P+P) approach. We propose to estimate the distance of the ball from the goal-plane through its parallax displacement between the two views with respect to the physical planar surface of the goal-plane. The performance of the approach has been determined on synthetic data obtained simulating different real contexts. Moreover the method has been tested also on real images acquired with a binocular system appropriately positioned in a real environment.