Video Refereeing Model of Soccer Match Based on Fuzzy Clustering and Cuckoo Optimization Algorithm

Abstract

Computer technology has rapidly advanced the use of computer-assisted refereeing in sports. However, motion image segmentation presents several challenges, including image blurring due to high-speed motion and the impact of ambient light on image accuracy. These challenges can affect the referee’s judgment of the game. To address this issue, the research utilizes sinusoidal mapping and the simplex method to enhance the cuckoo algorithm. Additionally, the optimization accuracy is improved through sinusoidal adaptive discovery probability. To tackle the problem of the fuzzy clustering algorithm’s weak global search ability, the clustering center is optimized by enhancing the cuckoo algorithm. To optimize the objective function, domain space information is introduced into the fuzzy clustering algorithm to address its sensitivity to noise. As the number of iterations approached 500, the experimental findings showed that the convergence accuracy of the particle swarm algorithm, the sparrow algorithm, the modified cuckoo algorithm, and the cuckoo algorithm were, respectively, 0.99, 0.89, 0.85, and 0.73. The enhanced algorithm model, enhanced cuckoo-fuzzy C-mean algorithm model, cuckoo-fuzzy C-mean algorithm model, and fuzzy C-mean algorithm model had delineation effects of 0.80, 0.72, 0.70, and 0.61, in that order. The outcomes show that the suggested algorithm can more accurately segment images from soccer matches. This provides a valuable reference for future computer-assisted match discrimination.

Keywords

• Image segmentation
• fuzzy clustering
• cuckoo’s algorithm
• match discrimination
• sine mapping