改进背景减法下人体运动模糊图像检测仿真  

Improve Simulation of Human Motion Fuzzy Image Detection under Background Subtraction

作　　者：靳铁柱 刘生彦 JIN Tie-zhu;LIU Sheng-yan(Xingzhi College,Xi'An University of Finance and Economics,Xi'an Shaanxi 710038,China)

机构地区：[1]西安财经大学行知学院,陕西西安710038

出　　处：《计算机仿真》2025年第3期304-308,共5页Computer Simulation

基　　金：陕西省“十四五”教育科学规划课题2023年度课题(SGH23Y2707);陕西省体育局2023年常规课题(2023356);西安财经大学行知学院2023年校级科研团队(大学体育与健康教育研究团队23KYTD06)。

摘　　要：人体在高速运动过程中,由于运动模糊效应和拍摄设备的限制,导致捕获的图像质量下降,难以准确识别运动目标。针对上述问题,提出改进背景减法的人体高速运动模糊图像检测。首先,通过对图像进行3*3分块处理,有效减少了计算复杂度并提升了算法效率。然后,基于混合高斯函数进行分块图像的背景模型的重建,突出了运动目标背景。接着,引入学习效率因子实现背景模型的自适应更新,避免动态背景干扰。最后,通过帧间差分法精确去除背景,分离出运动目标前景,并利用梯度相似度理论去除前景图像中的细微噪声,进一步增强了运动模糊图像前景目标清晰度。实验结果表明,所提方法在人体高速运动模糊图像检测中具有较高的准确性和鲁棒性,为实际应用提供了有力支持。In the process of high-speed movement of the human body,due to the motion blur effect and the limitation of shooting equipment,the quality of captured images decreases,and it is difficult to accurately identify moving targets.In order to solve this problem,an improved background subtraction method is proposed to detect the blurred image of human high-speed motion.Firstly,by processing the image into 3*3 blocks,the computational complexity is effectively reduced and the algorithm efficiency is improved.Then,the background model of the block image is reconstructed based on the mixed Gaussian function,which highlights the background of the moving target.Then,the learning efficiency factor is introduced to realize the adaptive update of the background model and avoid the dynamic background interference.Finally,the background is accurately removed by the inter-frame difference method,and the foreground of the moving object is separated,and the subtle noise in the foreground image is removed by the gradient similarity theory,which further enhances the clarity of the foreground object in the motion blurred image.The experimental results show that the proposed method has high accuracy and robustness in human high-speed motion blurred image detection,which provides strong support for practical application.

关 键 词：背景减法 图像处理 目标检测 背景更新 高斯模型 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]