Aiming at solving the problem of detecting the wideband chirp signals under low Signal-to-Noise Ratio (SNR)
condition, an effective signal detection algorithm based on Short-Time-Fourier-Transform (STFT) is proposed.
Considering the characteristic of dispersion of noise spectrum and concentration of chirp spectrum, STFT is performed
on chirp signals with Gauss window by fixed step, and these frequencies of peak spectrum obtained from every STFT are
in correspondence to the time of every stepped window. Then, the frequencies are binarized and the approach similar to
mnk method in time domain is used to detect the chirp pulse signal and determine the coarse starting time and ending
time. Finally, the data segments, where the former starting time and ending time locate, are subdivided into many
segments evenly, on which the STFT is implemented respectively. By that, the precise starting and ending time are
attained. Simulations shows that when the SNR is higher than -28dB, the detection probability is not less than 99% and
false alarm probability is zero, and also good estimation accuracy of starting and ending time is acquired. The algorithm
is easy to realize and surpasses FFT in computation when the width of STFT window and step length are selected
properly, so the presented algorithm has good engineering value.