The Optimized Method Of Multi-source Least-squares Reverse-time Migration

The technology of multi-source seismic acquisition allows more than one source to be activated at the different point,to obtain blended seismic data from multi-source.The technology has high efficiency and can reduce the acquisition cost effectively.Currently,imaging method of the multi-source seismic data can be divided into two kinds.The first method will separate the multi-source seismic data firstly,and then apply the conventional migration,while the second method will carry out imaging on multi-source seismic data directly.Among them,direct imaging is a commonly used method,which is efficient and can avoid the error of multi-source seismic data separation.But the disadvantage of this method is that it will introduce crosstalk noise caused by different gathers coherent into migration section,which will affect the imaging accuracy.The imaging result of Least-squares migration has better amplitude balance,better resolution,better fidelity and fewer migration artifacts than the result of conventional migration.So it can be applied to suppress the crosstalk noise caused by direct imaging of multi-source seismic data.But the conventional least-squares migration cannot quickly eliminate the influence of crosstalk noise on the imaging results to meet the requirements of imaging accuracy.Therefore,in order to ensure the stability of inversion and improve the inversion results,based on the realization of multi-source least-squares reverse-time migration and considering the sparseness of the reflection coefficient in Seislet domain,this paper first introduces Seislet transform to sparse constraint the final imaging results,and therefore realize the sparse constrained multi-source least-squares reverse-time migration process.The numerical experiments show that the method can effectively suppress the crosstalk noise,while it will damage the complex structure reflection of the event.Aiming at the problem,we develop a multi-source least-squares reverse-time migration method with the structure-oriented filtering.We introduce the structure-oriented filtering operator into the multi-source least-squares reverse-time migration as preconditioner,and can apply non-stationary filtering along the structural trend to suppress random noise under the premise of protecting structural information.The proposed method utilizes the conjugate-gradients method to minimize the mismatch between the predicted data and the observed data and obtains imaging result with higher signal to noise ratio.The numerical experiments show that the proposed method can effectively suppress the crosstalk noise and improve the imaging accuracy.