A Video Analyzing Method for Competitive Rifle Shooting

摘要

In this paper, we propose a video analysis method for scientific training of rifle shooting. Rifle target shooting is one of the Olympic games, it is a competition aiming at a 45.5mm target of 10m away. Due to the small number of players, computer-based training assistance has not been provided. When trying to analyze the movement of the rifle shooter’s form using video analysis, unnecessary movements are detected as noises. If the noise is removed by the conventional video processing filters, the feature points that are originally required will be removed also. In this paper, we propose a noise removal method using machine learning. Experimental results show that the proposed method has an accuracy of about 90%.