基于IQ图特征的通信辐射源个体识别

Specific Emitter Identification of Communication Radiation Source Based on the Characteristics IQ Graph Features

在真实的战场环境中,我们很难采集到足够的带标签的敌方辐射源数据,通过不断地发展,CNN神经网络有着很强的处理图片分类的能力,为了充分利用发展最为成熟的CNN神经网络,本文提出了一种将IQ路数据转化成图片的识别方法。由于数据的IQ图具有重复性与个体的差异性,通过实验,这种方法在识别不同个体超短波电台上有着94%的正确率,对比双谱特征,IQ图特征具有更强的识别能力。这种特征变换方法简单,并且CNN网络处理图片分类的技术成熟,具有很强的实用性。

In the real battlefield environment,it is difficult for us to collect enough labeled enemy radiation source data.Through continuous development,CNN neural network has a strong ability to process image classification,in order to make full use of the most mature CNN neural network.Network,this paper proposes a recognition method that converts IQ data into pictures.Because the IQ map of the data has repeatability and individual differences,through experiments,this method has a 94%correct rate in identifying different individual ultrashort wave radio stations.Compared with the bispectral feature,the IQ map feature has a stronger recognition ability.This feature transformation method is simple,and the CNN network processing image classification technology is mature and has strong practicability.