基于随机抽样一致性算法的运动平面上点的视觉测量及其应用  

RANSAC Based Visual Measurement of Point on Ground Plane and Application in Sports

作　　者：王岚[1] 沈乐君[1] 柯遵渝[1] 

机构地区：[1]成都体育学院,四川成都610041

出　　处：《中国体育科技》2008年第1期12-16,共5页China Sport Science and Technology

摘　　要：使用计算机视觉技术,对球场中运动员和关键点的位置进行测量,是基于视频的技、战术分析系统的核心技术,也是运动解析的重要原理,更是体育科学研究的重要工具。介绍其理论基础与实现方法,探讨测量误差和RANSAC容错算法,介绍使用三维模型进行标定和结果演示。实验证明,提出的方法和传统方法相比,提高了纠错能力和适应性,在体育科研中有广泛的应用价值。Computer vision based visual measurement of points on ground plane via video or digital image are introduced by paper. What we presented is fundamental for sport software and crucial for the tactic analysis. This paper introduces the basic theory, discusses the error of measurement and use RANSAC algorithm framework to improve algorithm accuracy. The 3D model can visible and very useful when we do calibration and presentation of analysis. The result of experiments showed that the method in the paper is more robust, flexible than traditional method and can be widely used in many fields of sport science.

关 键 词：视觉测量 RANSAC 单应性矩阵 图像坐标 球场坐标 击球落点 网球 

分 类 号：G80-05[文化科学—运动人体科学]