Abstract: When the infrasound wave signals from leakage of gas pipelines are collected, signal tips detected by infrasound wave sensors are abandoned due to the frequency domain limit of the sensor, while the ambient noise in the natural world produces a strong interference on the acoustic signal. To reduce false positive rates of infrasound wave detection, it is necessary to make a tip compensation for detection signal through a special algorithm, quickly separate the real leak signals from interference signals, and select the wavelet transform method to process signals after compensation to make collected signals have good local features in time domain and frequency domain and have different resolutions in different position of time-frequency plane, so as to form better inhibitory effect of the noise and more accurately extract signal mutation point. Through the above tip compensation and wavelet transform signal processing, false positive rates and false negative rates can be reduced more thoroughly.