Multifrequency Integration and Scale-Frequency Linear Attention for Aerial Tracking

Aerial tracking is an essential component of vision-based measurement and UAV systems, playing a crucial role in autonomous navigation, intelligent transportation, and remote sensing. However, aerial scenarios present unique challenges such as rapid motion, scale variation, and camera movement. Most existing tracking methods either use CNNs for feature extraction or rely on Transformers for global feature modeling, but these approaches struggle to balance accuracy and real-time performance. To address this issue, this paper proposes MISATrack, an efficient Siamese network model that integrates frequency domain encoder and scale-frequency linear attention mechanism. Firstly, we apply discrete wavelet transform to convert the input image into the frequency domain, and then enhance its diverse visual information by our multi-frequency integration strategy. In addition, we further design scale-frequency linear attention module for hierarchical feature fusion. This scheme greatly facilitates representation learning with few computation overhead, resulting in more robust object tracking. The proposed tracker is evaluated using four UAV tracking benchmarks: DTB70, UAV123, UAV123@10fps and UAVTrack112. Experimental results indicate that MISATrack outperforms most state-of-the-art trackers while maintaining real-time tracking. The code is publicly available at https://github.com/Wang123z/MISATrack.