UAV ground motion target tracking algorithm based on optical flow clustering

Aiming at the situation that it is difficult for UAVs to track the moving targets due to the ambient colour feature is not obvious, the data transmission is blocked and the computing power of the UAV platform is insufficient, a ground moving target tracking algorithm based on optical flow clustering is proposed, i.e. the speed calculated based on the position and the optical flow and the clustering of the feature points achieve the tracking of the moving targets. Through the verification of real blurred video, it is proved that the algorithm can accurately detect and track the ground moving targets with low misdetection rate, which is of some practical significance in practical applications.