Occupancy Analysis of Soccer Fields Using Wide-Angle Lens

Abstract

Occupancy analysis is critical for resource assessment. This paper presents a novel solution for occupancy analysis in soccer fields, which is needed to assist the management for resource assessment. The analysis is based on player detection in the soccer fields. The process of detection is performed by using one static fish-eye camera, which is achieved by enhancement of the players based on their structural properties using 2-D Gabor wavelet combined with background subtraction. Moreover, the gray scale intensity matching is performed for catering luminance issues. Occlusion is handled through a color and compactness based analysis of expected player regions. It has been shown through experiments that the developed method results in precise analysis of occupancy i.e. an average error of .0094 % for no player in the field, 2.67% for full activity in the field and 3.64% during transition.