The video target detection method of intelligent substation is studied based on improved GMM to accurately detect the video target of substation.A Gaussian distribution model is built for each pixel of the video sequence of the intelligent substa-tion,and to determine whether the pixel is the background by judging the matching between the pixel and the Gaussian distri-bution model.It can adjust the learning rate by the four neighborhood method,realize the weight adjustment of the first matc-hing Gaussian distribution model,and realize the descending arrangement of the background Gaussian distribution model ac-cording to the ratio of weight value to variance.It can also obtain the video background image of the intelligent substation,re-move the background image from the current frame image to obtain the foreground image,and detect the video target of the in-telligent substation by removing the discrete point noise,internal noise and shadow of the foreground image.The experimental results show that when the Gaussian distribution of pixels is 5,the video target effect in the foreground image of intelligent substation detected by this method is better,and the average values of false detection pixels and missed detection foreground pixels are lower.
关键词
改进GMM/智能变电站/视频目标检测/高斯分布模型
Key words
improved GMM/intelligent substation/video target detection/Gaussian distribution model
维普
万方数据