基于维纳滤波的体育运动视频模糊图像复原方法  

Restoration Method of Sports Video Blurred Image Based on Wiener Filter

作　　者：马滔[1] 刘秋梅 MA Tao;LIU Qiu-mei(School of Physical Education,Yan'an University,Yan'an 716000 China;Yan'an New Area,No.1 Primary School,Yan'an 716000 China)

机构地区：[1]延安大学体育学院,陕西延安716000 [2]延安市新区第一小学,陕西延安716000

出　　处：《自动化技术与应用》2023年第9期123-126,142,共5页Techniques of Automation and Applications

摘　　要：为解决由噪声、拍摄目标与摄像机间相对运动导致的视频成像质量下降的问题,提出一种基于维纳滤波的体育运动视频模糊图像复原方法。图像复原过程即为图像退化的逆过程,在复原前通过噪声及模糊图像构建退化模型。使用拉格朗日乘数法构建退化模型复原约束条件,计算维纳滤波函数中运动图像的模糊角度和模糊尺度,完成图像复原全过程。实验表明在不同的相对运动速度下,所提方法能够有效地去除模糊图像梯度效应,使图像中没有较明显的模糊点,能够基本满足体育运动对视频图像清晰度的要求。In order to solve the problem of image quality degradation caused by noise and relative motion between the target and the camera,a restoration method of sports video blurred image based on Wiener filter is proposed.The process of image restoration is the reverse process of image degradation,so the degradation model is constructed by noise and blurred image before restoration.The Lagrange multiplier method is used to construct the constraint conditions for restoration of degradation model,and the blur angle and blur scale of moving image in Wiener filter function are calculated to complete the whole process of image restoration.The experimental results show that the proposed method can effectively remove the gradient effect of blurred image under different relative motion speed,so that there are no obvious blurred points in the image,and it can basically meet the requirements of sports for video image clarity.

关 键 词：图像退化模型 拉格朗日乘数法 维纳滤波函数 模糊图像复原 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]