Enhancing the Scale Adaptation of Global Tracker for Infrared UAV Tracking

Tracking an unmanned aerial vehicle (UAV) in the infrared video is an essential technology for the anti-UAV task. Given the frequent UAV target disappearance caused by occlusion or moving out of view, the global tracker, which has the unique advantage of recapturing the target, is widely used in infrared UAV tracking. However, the global tracker performs poorly when dealing with large target scale variation, because it cannot maintain the approximate consistency between the target sizes in the template and the search region. To enhance the scale adaptation of global trackers, we propose a plug-and-play scale adaptation enhancement module (SAEM). It can generate a scale adaptation enhancement kernel according to the target size in the previous frame, and then perform implicit scale adjustment on the extracted target template features. To optimize the training, we introduce an auxiliary branch to supervise the learning of SAEM, and add Gaussian noise to the input size to improve the robustness of SAEM. In addition, we propose a one-stage anchor-free global tracker (OSGT), which has a more concise structure than other global trackers to meet the real-time requirement. Extensive experiments on three Anti-UAV Challenge datasets and the Anti-UAV410 dataset demonstrate the superior performance of our method, and verify that our proposed SAEM can effectively enhance the scale adaptation of existing global trackers.