利用数学形态学滤波方法压制多次波

Suppressing multiple by mathematical morphology filtering method

传统的基于波动方程的多次波处理流程分为预测多次波模型和自适应匹配相减两步,但是在多次波能量和有效波能量交叉时难以取得很好的效果.本文引入数学形态学滤波的方法解决这个问题.在获取了较高精度的多次波预测模型后,通过取极值法和相关法获取多次波同向轴形态,并用数学形态学滤波方法在原始地震数据上提取并分离多次波.数学形态学滤波是一种非线性滤波方法,其计算在时域内完成,具有较快的运算速度和良好的并行性.同其他滤波方法相比,在多次波同向轴和有效波同向轴相交时,可以取得较好效果.本文介绍了形态学滤波的基本原理,并通过简单模型、Pluto数据和实际数据的测试,验证了方法的有效性.

Traditional multiple suppression approach includes two steps：multiple model prediction and adaptive subtraction, but the results of this method are not satisfied enough when the energy of multiples and primaries are mixed. Mathematical morphology method is a useful tool in signal processing. We introduce mathematical morphology method to solve this problem. We first get a predicted multiple model, and then get the form of multiple＇s event by correlation and picking. Using the information of direction of events, we can extract the multiple from the original seismic dataset by mathematical morphology filter. This method is a non-liner method, all calculation are took in time domain. Comparing with other method, mathematical morphology filter can do a good job in the case of the events of multiples and primaries are intersected. This paper introduces the basic method of mathematical morphology. Through the processing of the model dataset, Pluto dataset and real seismic data, the validity of the proposed method is demonstrated.