“Researchers have proposed a twin neural network target tracking algorithm called SiamAU to address the issues of low altitude drone flight and small pixel ratio in video images. On the basis of SiamRPN++, this algorithm introduces an improved backbone network and feature rearrangement mechanism. By adding an ECA Net attention mechanism network and improving the activation function, it enhances the feature representation ability in complex backgrounds. At the same time, the algorithm also performs shallow dimensionality reduction and deep feature fusion to obtain improved deep fusion features that are more suitable for tracking small targets such as drones. Tests on the DUT Anti UAV dataset showed that the success rate and accuracy of the SiamAU algorithm reached 60.5% and 88.1%, respectively, representing improvements of 5.6% and 8.1% compared to the benchmark algorithm. This research achievement provides a new solution for target tracking in anti drone missions and is expected to open up new directions for research in related fields.”