针对传统时频峰值滤波(TFPF:Time-Frequency Peak Filtering)未考虑地震信号记录道与道之间的相关性以及径向道TFPF未考虑实际同相轴走向的问题,在假设反射同相轴局部线性的基础上,提出自适应径向道时频峰值滤波算法。该算法利用自适应...针对传统时频峰值滤波(TFPF:Time-Frequency Peak Filtering)未考虑地震信号记录道与道之间的相关性以及径向道TFPF未考虑实际同相轴走向的问题,在假设反射同相轴局部线性的基础上,提出自适应径向道时频峰值滤波算法。该算法利用自适应函数获取局部区域内同相轴的走向,通过径向道变换(RTT:Radial-Trace Transform)将同相轴拉伸到径向道域,提高其线性度,更好地满足TFPF无偏估计的条件,减小因信号非线性引起的估计误差。实验结果表明,该方法在随机噪声压制和有效信号恢复方面都取得了更好的效果。相同窗长下,该算法相比传统TFPF信噪比提高了5 d B左右,相比径向道TFPF信噪比提高了3 d B左右。展开更多
Ground roll waves interfere with seismic data. The suppression of ground roll waves based on the division of wavelet frequencies considers the low-frequency characteristics of ground roll waves. However, this method w...Ground roll waves interfere with seismic data. The suppression of ground roll waves based on the division of wavelet frequencies considers the low-frequency characteristics of ground roll waves. However, this method will not be effective when the ground roll wave and the effective signal have the same frequency bands because of overlapping. The radial trace transform （RTT） considers the apparent velocity difference between the effective signal and the ground roll wave to suppress the latter, but affects the low-frequency components of the former. This study proposes a ground roll wave suppression method by combining the wavelet frequency division and the RTT based on the difference between the ground roll wave velocity and the effective signal and their energy difference in the wavelet domain, thus making full use of the advantages of both methods. First, we decompose the seismic data into different frequency bands through wavelet transform. Second, the RTT and low-cut filtering are applied to the low-frequency band, where the ground roll waves are appearing. Third, we reconstruct the seismic record without ground roll waves by using the inverse RTT and the remaining frequency bands. The proposed method not only improves the ground roll wave suppression, but also protects the signal integrity. The numerical simulation and real seismic data processing results suggest that the proposed method has a strong ability to denoise while preserving the amplitude.展开更多
文摘针对传统时频峰值滤波(TFPF:Time-Frequency Peak Filtering)未考虑地震信号记录道与道之间的相关性以及径向道TFPF未考虑实际同相轴走向的问题,在假设反射同相轴局部线性的基础上,提出自适应径向道时频峰值滤波算法。该算法利用自适应函数获取局部区域内同相轴的走向,通过径向道变换(RTT:Radial-Trace Transform)将同相轴拉伸到径向道域,提高其线性度,更好地满足TFPF无偏估计的条件,减小因信号非线性引起的估计误差。实验结果表明,该方法在随机噪声压制和有效信号恢复方面都取得了更好的效果。相同窗长下,该算法相比传统TFPF信噪比提高了5 d B左右,相比径向道TFPF信噪比提高了3 d B左右。
基金supported by the National Science and Technology Major Project(No.2011ZX05007-006)the 973 Program of China(No.2013CB228604)the major Project of Petrochina(No.2014B-0610)
文摘Ground roll waves interfere with seismic data. The suppression of ground roll waves based on the division of wavelet frequencies considers the low-frequency characteristics of ground roll waves. However, this method will not be effective when the ground roll wave and the effective signal have the same frequency bands because of overlapping. The radial trace transform （RTT） considers the apparent velocity difference between the effective signal and the ground roll wave to suppress the latter, but affects the low-frequency components of the former. This study proposes a ground roll wave suppression method by combining the wavelet frequency division and the RTT based on the difference between the ground roll wave velocity and the effective signal and their energy difference in the wavelet domain, thus making full use of the advantages of both methods. First, we decompose the seismic data into different frequency bands through wavelet transform. Second, the RTT and low-cut filtering are applied to the low-frequency band, where the ground roll waves are appearing. Third, we reconstruct the seismic record without ground roll waves by using the inverse RTT and the remaining frequency bands. The proposed method not only improves the ground roll wave suppression, but also protects the signal integrity. The numerical simulation and real seismic data processing results suggest that the proposed method has a strong ability to denoise while preserving the amplitude.
