基于孪生网络的高效无人机目标跟踪  

Efficient Tracking for UAVs Based on Siamese Network

作　　者：王建浩 叶明[1] 姚佳烽 WANG Jian-Hao;YE Ming;YAO Jia-Feng(College of Mechanical&Electrical Engineering,Nanjing University of Aeronautics&Astronautics,Nanjing 210016,China)

机构地区：[1]南京航空航天大学机电学院,南京210016

出　　处：《计算机系统应用》2025年第1期69-79,共11页Computer Systems & Applications

基　　金：国家自然科学基金面上项目(62271251)。

摘　　要：在视觉跟踪领域,大多数基于深度学习的跟踪器过分地强调精度,而忽视了算法速度.因此,这些算法在移动平台上的部署(无人机),受到了阻碍.在本文中,提出了一种基于Siamese的深度交叉指导跟踪器(SiamDCG).为了更好地在边缘计算设备上部署,在MobileNetV3-small的基础上设计了独特的backbone结构.此外,针对无人机场景的复杂性,传统使用狄拉克δ分布预测目标框的方式有很大的弊端,为了克服边界框存在的模糊效应,Siam DCG将回归框分支转为预测偏移量的分布,并且用学习到的分布去指导分类的准确性.在多个无人机benchmark上的优秀表现,都显示了其鲁棒性与高效性.在Intel i512代CPU上,SiamDCG运行速度是SiamRPN++的167倍,使用的参数仅为它的1/98,FLOPs是1/410.In the field of visual tracking,most deep learning-based trackers overemphasize accuracy while overlooking efficiency,thereby hindering their deployment on mobile platforms such as drones.In this study,a deep cross guidance Siamese network(SiamDCG)is put forward.To better deploy on edge computing devices,a unique backbone structure based on MobileNetV3-small is devised.Given the complexity of drone scenarios,the traditional method of regressing target boxes using Diracδdistribution has significant drawbacks.To overcome the blurring effects inherent in bounding boxes,the regression branch is converted into predicting offset distribution,and the learned distribution is used to guide classification accuracy.Excellent performances on multiple aerial tracking benchmarks demonstrate the proposed approach’s robustness and efficiency.On an Intel i512th generation CPU,SiamDCG runs 167 times faster than SiamRPN++,while using 98 times fewer parameters and 410 times fewer FLOPs.

关 键 词：目标跟踪 轻量级网络 孪生网络 无人机 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]