基于机器视觉的真人射击游戏智能仲裁系统

Intelligent scoreboard system for wargame based on machine vision

为解决真人射击游戏得分和输赢难以评判的问题,提出一种基于机器视觉技术的射击游戏仲裁系统设计。在该系统中,游戏参与者身上贴有特殊设计的AprilTag靶标,游戏枪上配置智能图像识别装置,在击发时检测靶标与虚拟弹着点的误差,判定对手是否被击中。通信模块将判定结果上传到仲裁系统服务器网关进行信息汇总和发布,通过手机客户端访问网关,即可实时查看比赛得分及排名。在靶标识别范围的改进方面,使用了双层嵌套式标签的设计方案,支持游戏枪在远近距离处均能准确识别目标,有效检测范围为0.5~20 m。在靶标定位精度的提升方面,将贝叶斯数据融合算法应用于嵌套式靶标的位姿评估,该算法使靶标定位误差减小了10%。

To score and judge who is the winner in wargame, a wargame scoreboard system based on machine vision technology is presented.In this system, a specially designed AprilTag is affixed to the game players, and the gun is equipped with an intelligent camera, which detects the deviation between the target and the virtual impact point when the gun is triggered, so to estimate whether the player make an effective shot or not.Besides, the communication module uploads the result to the server for collecting and releasing information, and players can access gateway by application(App)on their smart phone.Double-layer embedded tags are designed for improving tag recognition range, which support the gun to detect the target at both long and short distances, with effective detecting range of 0.5~20 m.Bayesian data fusion algorithm is applied for pose estimation of the embedded tag, and the proposed algorithm reduces the position error by 10 %.