基于动态时间规整ICA算法地震随机噪声压制

噪声压制是地震数据处理流程中的基本环节之一。传统的独立分量分析(ICA)算法仅适用于平缓地层同相轴的地震资料噪声压制,对非平缓地层同相轴地震资料去噪效果较差,且算法不够稳定,容易出现解混失败现象,导致去噪结果中产生坏道。针对这些问题,提出了将ICA算法与动态时间规整(DTW)算法相结合的噪声压制方法。首先使用DTW算法将倾斜地层同相轴校正为水平同相轴,利用ICA算法提取拉平后含噪地震数据的独立分量,实现拉平地震道的信噪分离。然后利用由DTW算法所提取的道间时差将同相轴还原为倾斜地层同相轴,从而实现复杂地震资料的随机噪声压制。理论模型和叠前叠后实际地震资料测试结果表明,该方法可以有效地压制地震数据中的随机噪声,且对非平缓地层也有较好的去噪效果,具有一定的实用价值。

Noise suppression is the one of the key issues associated with seismic data processing.The traditional Independent Component Analysis (ICA) algorithm is only suitable for noise suppression in seismic data with flat events.For complex seismic data with non-flat events,the denoising ability of the traditional ICA algorithm is poor,as the algorithm is not stable enough,being prone to demixing failure,resulting in bad sectors in the denoising results.To solve these problems,a noise suppression method combining the ICA algorithm with a dynamic time warping (DTW) algorithm is proposed.Firstly,the DTW is used to flatten tilted seismic data events,and then the ICA algorithm is used to extract the independent component from the noisy data after flattened,and separate the signal from the noise.Next,the moveout between traces extracted by DTW algorithm can be used to restore the flattened events back to the original tilted events,so as to achieve random noise suppression of the complex seismic data.In tests on synthetic data,pre-stack and post-stack data showed that this method can effectively suppress random noise in seismic data,and that it performs well at denoising seismic data with complex non-flat events.

国家重点研发计划项目(2016YFC0601100)和国家自然科学基金项目(41304080)共同资助。