一种基于卷积神经网络的雷达目标分类方法

雷达作为对低空和地面目标探测及监视预警的主要手段,在安全领域应用广泛。针对现阶段实际应用中雷达目标分类技术中过于依赖人工提取特征的问题,提出了一种基于卷积神经网络的分类方法,对雷达回波数据进行二维傅里叶变换得到距离-多普勒图像,再以距离-多普勒图集作为数据集,训练神经网络,得到能够完成雷达目标识别的网络模型。结果表明,相较于传统方法,基于卷积神经网络的目标识别模型在省去人工工作的同时提高了目标识别精度。

Radar target classification based on convolutional neural network

As the main means of detecting and monitoring early warnings of low-altitude and ground targets,radar is widely used in the security field.Aiming at the problem of relying on artificial extraction features in radar target classification technology in practical application at present,a classification method based on convolutional neural network is proposed.The radar echo data is subjected to two-dimensional Fourier transform to obtain the distance-Doppler image,and then the distance-Doppler atlas is used as the data set to train the neural network,and the network model capable of radar target recognition is obtained.The results show that the target recognition model based on convolutional neural network improves the target recognition accuracy while eliminating manual work,compared with the traditional method.