基于波场分离理论的逆时偏移成像条件研究及应用

Study and application of imaging condition for reverse-time migration based on wave-fields separation

逆时偏移成像方法与其它算法的成像方法相比,由于算法不受地层倾角的限制,因此针对陡倾角构造和复杂地质模型地区的成像有着无可比拟的优越性,特别是在盐丘发育的地区优势更加明显.实际应用过程中,逆时偏移通常是应用互相关成像条件构建成像,然而,这种成像条件会产生低频、强振幅的噪音,如果不能彻底消除这些噪音,那它就会严重影响甚至淹没有效信号,特别是在浅层的强反射界面上方往往会出现大量的低频噪音,从而使得浅层构造基本无法识别.基于传统互相关成像条件的这一局限,本文提供了一个新的成像条件,它能在有效成像的同时消除这些低频、强振幅的干扰噪音;其具体实现思路就是将炮点及检波点波场分离成它们的单程波传播分量,然后采用互相关成像条件对分离后的波场进行成像,从而实现逆时偏移成像.通过对模型数据和实际野外数据的实测,证明了该方法的有效性和适用性.

Reverse-time migration （RTM） has more for imaging complex structure and steep dip area. Especially, it has obvious ad,antages in salt area, because it is without the dip limitation of travel operator. We often adopt correlation-based imaging condition in industrial production process. However, low-frequency and large-amplitude noises are commonly seen in the section after migration when the correlation -based imaging condition was in application, the noises can usually contaminate the signals in the image, and it often overrides the shadow structures, so it lead us can not distinguish the true structures, it is parasitical for us. Therefore, it is important to remove the noises form the results, but it is difficult to achieve, we generally use band-pass filtering to attenuate the low-frequency noises from section, but band-passfiltering can not remove the noises entirely. Based on this problem, many geopny attenuate these noises. And this paper provides a new imaging condition which can eliminate the noises, it also can obtain good imaging results same time. As we know, the result obtained from one-way equation migration has no noises like the result obtained from RTM, this event guide us to cause a new idea. The basic theory is based on wave-fields separation. First of all, we introduce the basic theory of low-frequency and high-amplitude noises in traditional RTM image then we describe the theory and workflow of new method, the first step we have to do is that we must decompose both of the source wave-fields and receiver wave-fields to their one-way propagation components, we can obtain four terms wave-fields, up-going source wave-fields, down-going source wave-fields, upgoing receiver wave-fields and down-going wave-fields, from what have we discussed above, we have knew the basic reason why the noises generated when correlation-based imaging condition was in application, it caused by the correlation between the source wave-fields and receiver wave-fields which have the same propagate direction. So we only need the correlation between the source wave-fields and receiver wave-fields which have the different direction. After this work, we use the correlation-based image condition to the appropriate combinations of the decomposed wave fields like traditional imaging condition. From what we have discussed above, we only need two correlations, the first correlation is between up-going source wave-fields and down-going receiver wave-fields, and the second correlation is between up-going receiver wave-fields and down-going source wave-fields. At the same time, we give an effective method to separate source wave-fields and receiver-fields to their one-way propagation components, we achieve this step using Fourier transform- However, this imaging condition asks us to separate the wave-fields, It increases the calculated amount, but it is inessential with the development of the hardware like GPU technique. In order to show the adaptability and stability of proved method, we use the new imaging condition to different synthetic data and field data. It shows that new imaging condition for RTM based on wave-fields separation can remove the noises from the RTM results entirely. It is effective and practicable.