Currently, some single object tracking algorithms have achieved leading performance, but their large models limit their application on resource-constrained platforms such as Unmanned Aerial Vehicles (UAVs). This paper designs a lightweight single object tracking algorithm for UAVs based on the Siamese network to achieve efficient tracking with less resource consumption. Firstly, a lightweight backbone network for Siamese feature extraction is designed based on MobileNetV3, significantly reducing the computationand parameter volume of the network without compromising feature extraction capabilities. Secondly, a dual cross-correlation module is designed. The module uses Pointwise Cross-Correlation to quickly calculate the similarity between template image featuresand search image featuresand also uses Depthwise Cross-Correlation to supplement missing features, effectively enhancing feature matching accuracyand robustness. Then, a lightweight prediction head is designed by stacking multiple depthwise separable convolution layers, obtaining accurate target representations with minimal performance consumption. Finally, classification ranking loss is introduced to traditional classificationand regression losses, enhancing the network's ability to learn target foregroundand suppress background interference, and further improving tracking performance. Comprehensive experiments show that the proposed algorithm achieves 82.1%, 81.2%, and 64.6% precisionand 63.4%, 61.8%, and 49.6% success rate on DTB70, UAV123, and UAV20L datasets, respectively, with only 5.3×10⁵ parametersand 1.1×10⁸ floating point operations. It achieves the performance comparable to state-of-the-art tracking algorithms, while having significantly fewer parametersand computational load, and it can run at the speeds above 100 fps, meeting the real-time requirements of UAV object tracking.

Aiming at the problem of tracking failure due to target deformation, flipping and occlusion in visual tracking, a template updating algorithm based on image structural similarity is proposed by dynamically updating the template to adapt to the changes of the target during tracking. The tracking feature enhancement module and segmentation feature enhancement module are also designed based on the SiamMask network. The tracking feature enhancement module consists of non-local operations and convolutional downsampling, which is used to establish contextual correlation, enhance the target features, suppress the background interference, improve the tracking robustness, and solve the feature attenuation problem due to the occlusion of the target. The segmentation feature enhancement module introduces the convolutional block attention module and deformable convolution to improve the network’s ability to capture channel and spatial features, adaptively learn the shape and contour information of the target, and enhance the network’s segmentation accuracy of the tracked target, which in turn improves the tracking accuracy. In comparison to the baseline SiamMask, experiments demonstrate that the proposed algorithm performs well and steadily in solving the aforementioned problems, improving the expected average overlap rate by 0.052, 0.053, and 0.025 and the robustness by 0.06, 0.079, and 0.156 on the VOT2016, VOT2018, and VOT2019 datasets, respectively. It also achieves a real-time speed of 91 frames per second on average.