Morse电报信号由于其具有低功率、传输距离远等特点,仍被广泛应用于军用和民用的短波通信中.目前对短波宽带Morse信号的侦察主要采用人工监察方式,随着短波通信环境的日益复杂和Morse信号快速侦察需求的日益迫切,对宽带Morse信号自动检测技术的研究变得十分必要.本文首先对宽带数据进行短时傅里叶变换及图像预处理,得到信号时频图.接着采用自适应阈值实现信号区域定位.然后对定位得到的有效信号时频图集进行特征参数提取.最后,进行分类器训练并用于宽带Morse信号的自动检测.实验结果表明:在宽带环境下,本文方法对Morse信号检测正确率可达95％以上,漏检率和虚警率分别低于5％和12％.算法具有较好的检测效果和应用价值.%Because of the low power,long transmission distance characteristics,Morse telegraph signal is still widely used in military and civilian HF communications.Currently,artificial surveillance is used to detect short-wave broadband Morse signals.With the increasing complexity of shortwave communication environment and urgent demand of Morse signal rapid reconnaissance,the research on broadband Morse signal automatic detection technique becomes necessary.In this paper,first of all,Short time Fourier transform and image preprocessing were used to analyze the broadband data.Then adaptive threshold was used to implement signal regional locating.After that,the characteristic parameters were extracted from the locating signal time-frequency spectrum.Finally,classifier was trained and used to automatically detect broadband Morse signal.Experimental results show that the correct rate of Morse signal detection was more than 95％,the missing rate and the false alarm rate were separately less than 5％ and 12％ in the broadband environment.The proposed algorithm has a better detection effect and application value.
展开▼
