足球运动员角球射门视频图像轨迹提取仿真

摘要

有效提取足球运动员角球射门视频图像轨迹,可更好的提高足球运动员射门精度.对射门视频图像的轨迹进行提取,需要对运动员角球射门轨迹进行追踪,构成足球目标的初始轨迹,完成对轨迹的提取.传统方法通过获取目标与视场中心的偏差信息,完成射门过程中足球目标的跟踪控制,但无法建立跟踪目标的初始轨迹,导致提取精度低.提出基于卡尔曼滤波的交叉换位下足球运动员角球射门视频图像轨迹提取方法.以粒子滤波作为射门轨迹跟踪框架,利用视觉特征来获得射门过程中每帧图像的候选球,结合卡尔曼滤波器提取和线性差值理论填充足球角球射门过程中每帧图像漏检的球的位置,有效地完成了对交叉换位下足球运动员角球射门视频图像轨迹提取.实验结果表明,所提方法能够较好的提取交叉换位下足球运动员角球射门过程中的运动轨迹,且提取精度高.%This article proposes a trajectory extraction method for video image of football players' corner shoot under coordinated transpositions based on Kalman filter.Firstly,the particle filter is used as frame of shoot trajectory trace and visual feature is used to obtain candidate ball of each frame image during shoot.Then,the ball position missed in each frame image is filled during corner shoot integrated with the Kalman filter extraction and linear difference theory.Finally,the trajectory extraction is completed effectively.Experimental results show that the method can extract movement locus preferably.It has high extraction precision.