针对气枪主动源监测中相位加权叠加数据绝对到时拾取可能存在较大误差的问题，提出基于台站之间叠加信号波形的相似性，利用台站对波形的互相关得到更准确到时差的策略。在此基础上，将台站对双差地震速度成像方法直接用于台站对到时差数据，以确定地下介质的速度异常。本文利用长江安徽段气枪实验在庐江临时地震台阵接收到的数据进行了测试，发现基于波形互相关得到的到时差数据比绝对到时数据更准确。最后，采用台站对双差地震成像的方法，得到了庐江台阵覆盖区域的地壳三维速度异常模型，继而确定了该地区较大尺度的地下结构分布。

To overcome the challenging problem of picking accurate absolute arrival times on the phase weighted stacking waveforms from repeating active air-gun sources，we propose to take advantage of waveform similarity of the stacked data on stations to determine accurate station-pair differential arrival times by waveform cross correlation. Subsequently they can then be used directly to image velocity structures using the newly developed station-pair double-difference tomography method. We tested this method with the data recorded on the Lujiang seismic network from the Yangtze River Active Source Experiment in Anhui. We found that we could get more accurate station-pair differential arrival times with the waveform cross-correlation method than the manually picked arrival times. With the station-pair double-difference tomography method，we obtained the three-dimensional velocity model of the crust around the Lujiang seismic array，and then determined the relatively large-scale geological structures of the Lujiang area.

国家自然科学基金(41474039)资助