Research On Key Frames Extraction Of Flying Robot For Electrical Tower Inspection Video

Electrical towers and electrical elements inspection is an effective way to prevent and troubleshoot power equipment failure, it can ensure security and stability of power system operation. However, the current electrical towers and electrical components inspection mainly rely on manual inspection, life safety of Inspection staff not only difficult to guarantee, but also low inspection efficiency and high cost, and it is easy to be affected by bad weather or human factors. Therefore the use of flying robot equipped with the necessary equipment, from manual autonomous power equipment inspection has important practical significance.With flying robot as carrier mode power equipment inspection can guarantee personnel safety and less affected by the weather, but thus creating redundant data. In order to effectively deal with the video data of overhead transmission lines inspection of flying robot autonomous recorded, we propose a key frame extraction method combined with geographic information,aerial aircraft data and parameters. Under the premise of the integrity of data collection, to reduce data redundancy, improve inspection efficiency.Firstly, a key frame preprocessing methods, by comparison of the data and mathematical operations, the inspection video is saved as an image frame from the video stream, and is divided into two parts, the candidate key frames and redundant data. To ensure extract key frames accuracy. Secondly, using the method of equidistant random sampling and manual sampling selection, established exclusively for a video sample collection,which can sum up the candidate set keyframes main content, as comparison model of similarity calculation of subsequent image. Thirdly, with a the law of cosines under multi-dimensional space, in the candidate select most similar key frame to exclusive sample frame as the final keyframe. For the key frames extracted by the algorithm that proposed in this article, put forward the index of power equipment inspection's video key frame evaluation, and then integrated existing video key frame extraction algorithm to evaluate the results of this article's algorithm. Finally, we carried out experiments to prove the feasibility and effectiveness of the method. Through the integration of experiments,we build the key frames extraction system of electrical tower in flying robot inspection video.