Automatic detection of Morse signals in wide band environment

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.