基于盲源分离的ICA去噪技术在裂缝预测中的应用

地震记录中不可避免地包含随机干扰信号，直接利用原始的地震资料开展裂缝发育带预测，对裂缝预测中采用的边缘检测算法影响很大，降低了预测结果的准确性。因此，有必要对地震资料进行去除噪声的处理，提高原始地震资料的品质。本文基于盲源分离的独立分量分析方法(Independent component analysis,ICA)去噪技术，将地震资料分解为不同级次的背景与储层目标反射响应，实现有效信号与随机噪声的区分，去噪效果优于常规去噪算法的效果，保证了去噪后有效信息基本不受损失，处理后地震资料横向波形特征的稳定性得到了较好改善。实际工区应用效果表明，利用去噪处理后的地震资料开展边缘检测裂缝预测，裂缝发育区分布规律与区域断裂发育特征具有较好的一致性，且与钻井揭示的裂缝发育特征吻合性较好，从而提高了火成岩裂缝发育区预测的可靠性。

Application of ICA Denoising Based on Blind Source Separation in Fracture Prediction

The random interfering noise contained in seismic record, if not removed properly, will inevitably pose a threat on the fracture development zone prediction accuracy because of greatly disturbing the key edge detection algorithm used in predicting step. Therefore, it is necessary to remove noise from seismic data and improve the quality of original seismic data. In this study, the independent component analysis(ICA) denoising technique, a blind source separation method, is used to decompose the seismic data into different levels of background and target reflection response of reservoir, and effectively make a distinction between effective signal and the random noise,which makes the processing result better than the conventional denoising algorithm. The processing ensures that the signal information basically does not suffer any losses and proffers a better lateral consistency in waveform characteristics. The field results show that, by applying the denoising method to the seismic data before edge detection, a robust fracture prediction result of fracture development zone distribution is achieved corresponding to the regional characteristics of fracture development, and it is in accordance with the drilling results of fracture development characteristics. This study improves the reliability of the fracture prediction of igneous rock zones.