主动源与被动源地震数据插值及联合数据成像

本文提出了两种情况下主动源数据和被动源数据的插值方法,并研究了两种数据在偏移成像中的互补效果.基于互相关法被动源数据重构原理,本文提出了结合多域迭代去噪技术的重构方法.提出了两种时间域主动源和被动源数据的插值方法,分别是共炮点域能量匹配插值和共检波点域最小平方匹配插值.然后对获得的主动源和被动源联合地震数据进行叠前深度偏移成像.在被动源活跃度不是很高的地区进行被动源地震勘探时,少量的主动源地震数据可以有效控制和补充被动源数据的成像效果.在稀疏炮点的主动源勘探中,有效利用被动源的信息能够在成像中增加更多的细节信息,提高成像质量.

Interpolation of seismic data from active and passive sources and their joint migration imaging

In seismic acquisition, we may only acquire a limited amount of data or lose parts of data because of the environment, cost or efficiency. This will inevitably influence the quality of final imaging results. In data processing, we usually use interpolation to solve the problem of sparse data. The conventional interpolation methods are usually based on mathematic principles, and compensate data after calculation. The passive sources are actually located in the subsurface. The passive seismic waves transmit to the surface according to the same principles as active seismic waves. As a result, they contain abundant real information about the subsurface. We consider compensating active source data with passive source data using interpolation methods. In addition, in the areas suitable for passive seismic exploration, when there are not enough sources, the quality of reconstruction is low. So we also consider compensating passive source data with active source data using interpolation methods.As the SNR of passive retrieved data reconstructed by the conventional method is low, the quality of imaging is influenced seriously. Based on conventional methods, we propose a multidomain iterative denoising reconstruction method, which is based on the Curvelet threshold iterative denoising and multilevel median filter. As to the two cases mentioned above, and considering the matching degree of passive virtual wavelets and active wavelets, we suggest two time-domain interpolation methods using active and passive seismic data. They are the energy matching interpolation in the common source domain and least square matching interpolation in the common receiver domain. When the two kinds of wavelets match well, we adopt the energy matching method to calculate the interpolation factors. When the two kinds of wavelets do not match well, we adopt the least square matching interpolation and linear interpolation methods to calculate matching factors. Using the above reconstruction and interpolation methods, we analyze three cases and compare the results before and after pre-stack depth migration.For the ideal case, we adopt 1000 random noise sources, which are distributed uniformly both in axial and lateral directions. The recording time is 1200 s. After reconstruction, we conduct pre-stack depth migration with active and passive source data. The results show that in the ideal case, there is a large amount of effective reflection information. In the imaging results, the imaging result of passive-source seismic data is very close to that of active-source data. The complex structures and reservoirs are displayed clearly. For the case with inadequate passive sources, we adopt 400 locally distributed noise sources. We estimate the source region by comparing the causal and un-causal parts of reconstruction results. Then we conduct interpolation and migration. We can see from the results that the deep detailed information is clearer after interpolation. For the case of sparse data in active-source seismic exploration, we use a few noise sources located at depth. From the interpolation results, we can see that the active-source data and passive-source data match well after matching. The imaging results show that the complex structures and deep information become clearer after interpolation.The numerical results show that effective use of active- and passive-source data can greatly improve the quality of imaging. The multidomain iterative denoising reconstruction method can suppress noises effectively. By adjusting the threshold value and filter windows, we can extract effective signals gradually. In the interpolation processing of active- and passive-source data, energy matching interpolation in the common source domain and least square matching interpolation in the common receiver domain can obtain good results in different cases. In the area with a few passive sources, we can estimate the source region from the causal and un-causal part of data. Adding a few active data in the quiet area can compensate the imaging results effectively. In the sparse shot active seismic exploration, the use of passive data can add some detailed information.