基于卷积神经网络的反无人机系统图像识别方法

针对无人机的无证飞行和随意飞行严重影响和威胁公共安全的问题，提出了反无人机系统。识别无人机是反无人机系统实现的关键之一，为此提出了一种基于卷积神经网络的图像识别无人机方法。运用自制光学系统采集设备采集了不同型号的无人机图片以及鸟类图片，设计了针对无人机小样本识别的卷积神经网络和支持向量机。运用设计的卷积神经网络分别对MNIST数据集、无人机图片以及鸟的图片进行了识别，同时也运用支持向量机识别无人机和鸟的图片，进行了对比实验。实验结果表明，设计的卷积神经网络在MNIST数据集上识别准确率为91.3%，识别无人机准确率为95.9%，支持向量机识别准确率为88.4%。对比实验表明，提出的方法可以识别无人机和鸟以及不同类型的无人机并且识别结果优于支持向量机，可用于反无人机系统识别无人机，给同类研究提供了借鉴。

Image recognition method of anti UAV system based on convolutional neural network

In view of the serious impact and threat to public security of UAV's undocumented flight and random flight, an anti UAV system was proposed. Recognition of UAV is one of the key points in the realization of anti UAV system. An image recognition method based on convolutional neural network was proposed. The self-made optical system was used to collect images of different UAVs and birds, and convolutional neural network and support vector machine for UAV small sample recognition were designed. The convolution neural network was used to identify MNIST data set, UAV image and bird image respectively. At the same time, support vector machine was used to identify UAV and bird image, and the experiment was carried out. The experimental results show that the recognition accuracy of the convolutional neural network is 91.3% in MNIST data set, 95.9% in UAV recognition and 88.4% in support vector machine (SVM). The experimental results show that the proposed method can identify UAVs, birds and different types of UAVs, and the recognition result is better than that of SVM. It can be used for the identification of UAVs in anti UAV system, which provides reference for similar research.