UAV Target Tracking Algorithm Based on Spatial-Temporal Aware Correlation Filters

Aiming at the problem that the linear updating mode of target tracking filter template will lead to low accuracy and template drift in long-term target tracking, a UAV target tracking algorithm based on spatial-temporal aware correlation filter is proposed. Firstly, in the process of filter training, the update of filter template is adjusted by adding a time regularization term, to enhance the sensing ability of the algorithm to the time sequence. Then, the space perception ability of the algorithm is enhanced by expanding the search area, and more effective training samples are obtained by means of cutting matrix. Finally, compared with the classical BACF algorithm, the simulation results show that the accuracy and precision of the proposed algorithm are greatly improved, which effectively verifies the robustness and performance of the proposed algorithm.