With the rapid development of image processing and computer vision technologies, extracting structural information from image sequences and analyzing their dynamic changes has become a research focus in multiple fields. Addressing the research needs of long air gap discharge phenomena, this paper proposes a topology structure extraction and dynamic trajectory analysis method based on image sequences. Through steps such as image preprocessing, skeleton extraction, branch identification, endpoint tracking, and spatiotemporal data association, it achieves automatic extraction and classification of discharge channels like streamers and leaders during long air gap discharges. This method demonstrates good robustness and scalability in both synthetic and real discharge image sequences, providing an effective technical means for studying the physical mechanisms of long air gap discharges.