With the continuous development of countermeasure and anti-countermeasure technology as well as the increasing application of low intercept signals, traditional signal extraction methods are no longer to satisfy user's needs for extraction of low intercept signal waveforms. In this paper, based on the discrete prolate spheroidal sequence (DPSS), the time-window function is used to analyze multiple linear frequency modulation signals. A method employing horizontal threshold decision probability to capture signal is proposed, and extraction of signal waveform is completed by adaptive binarization and improved Hough transform (HT). Simulation shows that the proposed method is better than the short-time Fourier transform algorithm in terms of noise control and can accurately extract the waveform at low SNR.

discrete prolate spheroidal sequence  /  Hough transform  /  multiple linear frequency modulation  /  short-time Fourier transform