基于高斯混合模型的电力仓库视频运动目标检测

Video Moving Target Detection in Power Warehouse Based on Gaussian Mixture Model

在电力仓库视频监控方面,存在着运动目标的各种行为,目标物体检测对发现异常行为具有十分重要的应用价值。为有效监测电力仓库中的运动目标,论文利用高斯混合模型对电力仓库视频开展运动目标物体检测,通过模型构建、参数更新和模型生成等步骤实现目标物体的检测,并通过Matlab软件对仓库视频进行实验仿真。实验证明高斯混合模型检测结果优于其他方法,可以较为完整、准确地检测电力仓库运动目标。

In the video surveillance of power warehouses,there are various behaviors of moving targets.The detection of moving targets has important application value for discovering abnormal behaviors.In order to effectively monitor the moving targets in the power warehouse,the Gaussian mixture model is used to detect the moving target video.The moving target detection is carried out through the steps of background model construction,parameter updating and background modeling,and the warehouse video is simulated experimentally by Matlab software.Experimental results show that the Gaussian mixture model is better than the frame difference method and optical flow method,and can detect moving targets of power warehouse completely and accurately.