Research On Video Summary Technology Of Transmission Line Inspection Video

With the gradual expansion of power grid scale,more and more substation equipment in transmission lines.In order to accurately know the operation status of transmission equipment,it is necessary to obtain a large number of transmission line inspection videos in the process of overhaul.Video summary of transmission line inspection video construction can quickly understand the main content of the video,so as to determine whether further research is needed.Therefore,video summary of transmission line inspection video construction has great research value in power.The principle of video summarization technology is to ensure not only representativeness but also integrity,but also to make the video summary closest to human subjective perception.In this paper,transmission line inspection video is taken as the research object.According to the problems faced by transmission line inspection video summarization technology,three key frame selection methods are studied.The main work accomplished in this paper is as follows:Video summarization method based on content minimal distortion principle is studied to construct video summary for transmission line inspection video.Firstly,the theoretical basis of the minimal distortion principle summarization method is introduced,including video shot detection and minimal distortion principle theory.Finally,experiments are carried out for transmission line inspection video,and the experimental results are analyzed.Video frames in the results of video summarization obtained by traditional minimum distortion method contain only video images with background of white clouds and farmland but not transmission line targets,which violates the principle of video summarization for transmission line inspection.Aiming at the problem that video frames contain less target information but are also selected as video key frames,RGB saliency algorithm is used to calculate the image,which has a good effect on the removal of pure background image in video.Aiming at the characteristics of targets in transmission line inspection video frames,the LBP(Local Binary Pattern)feature extraction is further carried out,and the Fisher classifier is trained by using LBP feature.By improving the LBP operator,the classification accuracy of the Fisher classifier is greatly improved,the classification efficiency of the image is improved,and the video summary is further improved.Content,summary efficiency and summary accuracy.In this paper,the key frames of video summary are extracted by using minimum distortion algorithm,image saliency algorithm and key frame selection algorithmbased on LBP,and the transmission line video summary containing only transmission line equipment targets is obtained successfully under the condition that the content is complete.