Time-frequency domain correlation method based on the synchrosqueezed wavelet transform for tunnel vibroseis and its application

Abstract

To meet the demand for long-distance and highly accurate detection in tunnel construction, a vibroseis source has been introduced into tunnel forward prospecting. However, serious noise in tunnels reduces the resolution of vibroseis signals, and small-scale structure recognition in tunnels requires higher signal resolution. Thus, we develop a time-frequency domain correlation method based on the synchrosqueezed wavelet transform for tunnel vibroseis data. Time-frequency domain correlation may capture more information than a single time or frequency domain correlation method, and its effectiveness depends on the time-frequency transformation method. The synchrosqueezed wavelet transform allows us to obtain high-resolution time-frequency spectra and thus helps to extract high-resolution effective signals. The advantages and disadvantages of different correlation methods are highlighted using numerical examples, in which the high resolution and strong robustness of the time-frequency domain correlation method based on the synchrosqueezed wavelet transform are demonstrated. A detailed field case study in an iron mine tunnel further demonstrates the reliability of the time-frequency domain correlation method based on the synchrosqueezed wavelet transform for practical data.