Fusion Navigation Information For Power Line Inspection Video Abstraction

As the continuous development of robot technology, more and more flying robots are used for Overhead Power line Inspection. The inspection data is always store in format of video. How to get the content of the video quickly is very important to improve the efficiency of the investigation of the failure. So the study of power line inspection video abstraction is necessary.This paper introduces a method of video abstraction for overhead power lines inspection videos based on navigation information. By providing video index for overhead power lines inspection videos, users can query and browse them more conveniently so as to speed up the investigation of the failure and improve efficiency of maintenance. In our method, navigation information is adopted to segment video scenes and filter the frames which do not contain electrical equipment. Compare to detect electrical equipment from image texture level, our method is more efficient and general. Clustering method is used to remove the redundant sequences of successive frames, so that the frames closest to the cluster center are chosen as the key frames. The experiment results show that the video has been compressed. The video index contains the key electrical equipment.