Alpine Skiing Tracking Method Based on Deep Learning and Correlation Filter

In outdoor sports such as alpine skiing, athletes hope to obtain information movement trajectory and speed achieve scientific training. The portable sensor has a sense of restraint affects the athlete’s performance. Cameras have difficulty capturing whole race in an environment. For skiing needs, we designed acquisition system, using unmanned aerial vehicle (UAV) ground cameras on path race. environment field, high-speed curvilinear flight characteristics UAVs cause challenges detection tracking by UAVs. To realize UAVs, propose algorithm combining neural networks correlation filters. A confidence score based motion distance interval frame is defined. When network detector less than threshold, filter used expand recognition range redetection. addition, created annotated fresh dataset. We compare our with four advanced algorithms UAV123 Three methods are evaluate performance method sequence From simulation results, outperforms comparison terms accuracy robustness. Therefore, application value training skiers.