A Key Frame Extraction Method of HEVC Video Based on Clustering Algorithm for Electric Utilities Management

It is necessary to enhance the efficiency and convenience of key frame extraction for power lines and equipment inspection. As more and more Unmanned Aerial Vehicles (UAV) equipped with camera are applied for power asset management, High Efficiency Video Coding (HEVC) standard is one of the video compression standards adopted by UAV-based inspection. Video data acquired from sports camera can be automatically processed according to the preprocessing and key frame extraction algorithms to enhance the efficiency. The video encoding and decoding is tested aiming at the aspen_1080p video and customized electric utilities video according to HEVC standard. According to the decoded YUV420 video frames and the Peak Signal-to-noise Ratio of frames, perform frame association analysis with the ISODATA clustering algorithm. The nearest 5 frames compared with the cluster center are regarded as the key frames. The number of key frames can be customized in practice. The clustering and key extraction results validate the key frame extraction method and provide more convenience and higher efficiency for electric utilities inspection and management.