Research on target tracking of UAV based on improved Camshift algorithm

When UAV tracks the ground target, the classicCamshift algorithm is susceptible to similar color background/target, occlusion, etc. The target tracking method based on improved Camshift is proposed A tracking template based on three-dimensional joint histogram is established by extracting the Hue, Saturation and LBP texture features of the tracking target, and the weighting values of the three feature components are adjusted by the adaptive weighting strategy to improve the tracking accuracy of the algorithm. When the tracking target is blocked, Kalman filtering mechanism is introduced to enhance the robustness of the algorithm. The experimental results show that the improved algorithm can meet the requirementsoftargettrackingaccuracyandreal-timefor UAV.