一种噪声环境下的雷达目标高分辨率距离像鲁棒识别方法

随着深度学习技术被应用于雷达目标识别领域，其自动提取目标特征的特性大大提高了识别的准确率和鲁棒性，但噪声环境下的鲁棒性有待进一步研究。该文提出了一种在噪声环境下基于卷积神经网络(CNN)的雷达高分辨率距离像(HRRP)数据识别方法，通过增强训练集和使用残差块、inception结构和降噪自编码层增强网络结构，实现了在较宽信噪比范围下的较高识别率，其中在信噪比为0 dB的瑞利噪声条件下，识别率达到96.14%，并分析了网络结构和噪声类型对结果的影响。 

Robust High Resolution Range Profile Recognition Method for Radar Targets in Noisy Environments

With the application of deep learning technology in the radar target recognition field, the automatic extraction of the target feature greatly improves the accuracy and robustness of the recognition, but its robustness in noisy environments needs to be further investigated. This paper proposes a robust target recognition method for radar High Resolution Range Profile (HRRP) data based on Convolutional Neural Networks (CNN). By enhancing training set and using the residual block, inception structure, and denoising sparse autoencoder layer to enhance the network structure, a higher recognition rate is achieved in a wider SNR range, under the condition of 0 dB Rayleigh noise, the recognition rate reaches 96.14%, and the influence of the network structure and noise type on results is analyzed.