面向智能视频监控中的运动目标检测方法研究
    点此下载全文
引用本文:方 勇,张建军.面向智能视频监控中的运动目标检测方法研究[J].计算技术与自动化,2022,(4):79-83
摘要点击次数: 266
全文下载次数: 0
作者单位
方 勇,张建军 (国网兰州供电公司甘肃 兰州 730050) 
中文摘要:运动目标检测是实现智能视频监控的基础,针对当前运动目标检测方法在复杂场景中适应性差的问题,提出了一种结合时空马尔可夫随机场模型和高斯混合模型的运动目标检测方法。在训练时空马尔可夫随机场模型时,采用高斯混合模型的参数更新算法计算邻域图像分割区域的均值和方差,并通过时空邻域标记场设置势函数。通过与传统目标检测方法的仿真比较,验证了该方法的优越性。结果表明,与传统的目标检测方法相比,该方法在复杂场景下具有更高的检测精度,能够更清晰地分割前景中的运动目标。
中文关键词:智能视频监控  运动目标检测  时空马尔可夫随机场模型  高斯混合模型  复杂场景
 
Research on Moving Target Detection Method for Intelligent Video Surveillance
Abstract:Moving target detection is the basis of intelligent video surveillance. Aiming at the problem of poor adaptability of current moving target detection methods in complex scenes, a moving target detection method combining spatiotemporal Markov random field model and Gaussian mixture model is proposed.When training spatiotemporal Markov random field model, the parameter update algorithm of Gaussian mixture model is used to calculate the mean and variance of neighborhood image segmentation region, and the potential function is set through spatiotemporal neighborhood marker field.The superiority of this method is verified by simulation comparison with the traditional target detection method.The results show that compared with the traditional target detection methods, this method has higher detection accuracy in complex scenes and can segment the moving targets in the foreground more clearly.
keywords:intelligent video monitoring  moving target detection  spatial temporal Markov random field  Gaussian mixture model  complex scene
查看全文   查看/发表评论   下载pdf阅读器