中文摘要：

提高勘探资料信噪比是地震勘探的主要内容之一.本项研究利用短时能量将记录分成能量均衡的若干段,获得减少尺度变换误差的分段时频峰值滤波方法,并运用端点拓展和与地震信号特征匹配的多级时窗参数改进时频峰值滤波精度.理论模型和共炮点资料处理结果表明,分段时频峰值滤波很好地消除了尺度变化误差引起的信号波形阶梯状畸变,能够在压制强随机噪声的同时更好地保留信号特征.对低信噪比勘探资料处理具有较大的应用价值.

英文摘要：

It is one of the major subjects of seismic exploration to improve signal to noise ratio of seismic data. The segmented time-frequency peak filtering （S-TFPF） is proposed to reduce the error induced by scale transform of TFPF by dividing a trace of seismic data into some segments with even energy based on the short-time energy principle. Moreover, endpoint delay method is utilized to avoid the endpoint distortion. The multiple time-windowing parameters, which match the characteristics of signal, are also set for the segments. The S-TFPF is applied to synthetic records and common-shot records, and the results show that the method can retain signal characteristics better while suppress the random noise. Segmenting can effectively attenuate distortion of signal caused by scale transform error and can be applied to practical seismic exploration at low SNR.