基于对角切片特征提取和深度学习的辐射源识别

针对复杂电磁环境下辐射源识别率低的问题，提出基于对角切片特征和深度学习的辐射源识别算法。利用辐射源信号双谱的个体特性，提取信号双谱对角切片特征作为深度学习模型的输入数据，采用Softmax分类器进行辐射源识别。仿真实验利用两部同型辐射源进行测试，结果表明该算法能识别个体辐射源，在低信噪比条件下也能获得高的辐射源识别率；相比于其他识别算法，双谱对角切片特征有更鲁棒的分辨性。

Emitter recognition based on diagonal slice feature and deep learning

For the problem of low recognition rate of emitter in complex electromagnetic environment,an algorithm based on diagonal slice feature and deep learning is proposed.The algorithm takes advantage of the individual characteristics of the bi-spectrum of the emitter signal,extracts the signal bi-spectrum diagonal slice feature as the input data of the deep learning model,and uses the Softmax classifier to recognize the emitter.The simulation results show that the algorithm can recognize the individual emitter and obtain high recognition rate of emitter under low signal-to-noise ratio,compared with those of other recognition algorithms,bi-spectral diagonal slice features have more robust resolution.