基于低频信息补偿的数据驱动Marchenko成像

基于常规地震干涉法和地震超越干涉法，提出了SI和BSI的结合方法SIBSI，即在SI被动源低频信息提取的基础上，重构主动源BSI地震数据，并利用BSI进行格林函数重构和面向目标的Marchenko成像.研究了基于频率优势的主动源低频重构方法，在完整保留了主动源信号高频信息的基础上，有效重构了低频信息，拓宽了地震数据的频带范围.讨论了含有自由表面多次波的地震数据在Marchenko成像中应用的方法.设计了一个含有高阻抗地层的模型，在该模型上使用SI低频信息重构BSI主动源地震数据，最后与纯主动源地震数据的格林函数重构和Marchenko成像进行了对比，证明了本文所提出方法的有效性、抗噪性以及在提高成像效果中的优势.

Low frequency information compensation based data-driven Marchenko imaging

In seismic exploration, it is often tried to avoid receiving ambient noise, but this situation doesn't suit seismic interferometry. We believe recorded passive seismic data have all kinds of information, both useful and useless for seismic imaging.#br#Both higher frequency and low frequency information are crucial for accurate seismic imaging. In this paper, we present the first successful application combining Seismic Interferometry with Beyond Seismic Interferometry for Green's function retrieval and Marchenko imaging, which is to use the low frequency part of the SI result to rebuild the active shot data's low frequency part.#br#The obstacle between SI and BSI data is their hypotheses, SI needs a free surface on the surface to fulfill the hypothesis of receivers surrounded by random sources, which means the free surface multiples are present in the final SI result. Based on matched filtering method, we use the low frequency part of SI to rebuild the low frequency part of BSI active data, this method is called "Low Frequency Advantages Based Low Frequency Information Retrieval". Before we use the merged broadband seismic data into Green's functions retrieval and Marchenko imaging, we proposed a data pre-processing method about the direct arrival and free surface multiple elimination of the merged reflection data, we use this method to get the reflection response which is needed by BSI Green's function retrieval. We discussed our method in a detailed way, we illustrate our results for showing the possibility to combine SI with BSI method, eventually, the reflectors are better focused and no more obvious artifacts are introduced which gives us a better and enhanced Marchenko imaging result.