【摘要】 随着无人机技术的快速发展,无人机技术已经在各个方面得到了广泛的应用,人们也越来越多的重视智能无人机技术的研究,其中如何提高智能无人机的自主循迹能力已经成为众多学者的研究重点。为了使无人机能够自主循迹飞行,提出了一种基于改进卷积神经网络的无人机循迹方法,在LeNet-5模型的基础上对原模型的层数进行改进,减小其复杂程度,使其可以运行在机载设备上,另外还给出了用于模型训练的图像数据的采集方法。其通过无人机上的相机采集前方图像,然后训练好的卷积神经网络会对此图像进行识别,并产生控制指令。实验结果表明,该方法可以有效地使无人机在楼道里自主的循迹飞行。更多还原

【Abstract】 With the rapid development of technology,unmanned aerial vehicle(UAV) technology has been widely applied in various areas,people are paying more and more attention to the research of intelligent technology,how to improve the ability of the autonomous tracking of intelligent unmanned aerial vehicle(UAV) has become a research emphasis of many scholars.This paper proposes a UAV tracking method based on improved convolution neural network,the number of layer of LeNet-5 model has been change to reduce the complexity,enable it to be run on the airborne equipment.The front camera on the unmanned aerial vehicle(UAV) will gathering images,and then the trained convolutional neural network will identify the image and generate control instruction. The experimental results showed that the method can effectively enable the UAV to autonomous flight in the corridor.更多还原