KCF-Match Target Tracking Algorithm for Tracking Swing Angle of Coupler Based on Video

The coupler is an essential component on the train that has the function of connecting and buffering. The actual dynamic performance of the coupler directly influences the safety and comfort of the vehicle. When the heavy haul train passes through the curve, the extreme swing angles of the couplers will seriously threaten the safety of the train. Therefore, the kernelized correlation filter-template matching (KCF-Match) target tracking algorithm is proposed to track the position and calculate the swing angles of the couplers. After the tracked area is selected, the corresponding data of the area are input into the KCF target tracking model for tracking. During the tracking process, if the tracking effects are not satisfied with the given evaluation indexes, the template matching algorithm will be used to track again. Experiments show that KCF-Match target tracking algorithm can achieve 99.8% accuracy rate and 99.9% success rate on the premise of ensuring real-time performance.