论文部分内容阅读
地震数据重构是地震数据处理的重要步骤之一,重构算法的精度、效率与抗噪性是地震数据重构技术的核心研究内容。研究针对傅里叶域凸集投影(POCS)算法,在定义的最优阈值评价标准基础上,提出了反比例阈值模型,该模型具有在大系数区间比指数模型更快下降速率、而在小系数区间比指数模型更慢下降速率,从而在保证弱反射信号重构精度的同时有效提高POCS地震数据重构算法计算效率。为提高反比例阈值对不同地震数据特点的适应性,在地震数据谱能量分布差异性特征分析基础上,研究提出了在反比例阈值模型分母上增加适应地震数据谱能量特征的因变参数,通过调节该因变参数获得适应不同地震数据特点的最佳阈值曲线,进一步提高算法的计算精度与计算效率。为了实现重构过程中随机噪音的自适应衰减,提高重构后地震数据信噪比,研究提出了数据驱动的加权回加系数计算策略,利用每次迭代对应数据驱动阈值占阈值区间的百分比获得加权回加系数。研究将新方法应用于模拟三维数据和实际三维地震数据,分析结果表明反比例阈值相对传统阈值在提高数据重构计算效率和精度方面具有明显的优越性,新提出的加权回加系数计算策略能有效提高重构数据的信噪比。
Seismic data reconstruction is one of the important steps in seismic data processing. The accuracy, efficiency and anti-noise of reconstruction algorithm are the core research contents of seismic data reconstruction. In this paper, based on the definition of the optimal threshold evaluation criterion, aiming at the Fourier domain convex projection (POCS) algorithm, an inverse proportion threshold model is proposed, which has a faster decline rate than the exponential model in the large coefficient interval, The interval is slower than the exponential model, which can effectively improve the computational efficiency of POCS seismic data reconstruction algorithm while ensuring the reconstruction accuracy of weakly reflected signals. In order to improve the adaptability of inverse proportion thresholds to the characteristics of different seismic data, based on the analysis of the characteristics of the energy distribution of the seismic data spectrum, this paper proposed a method of adding the denominator parameter to the denominator of the inverse proportion threshold model and adjusting the energy characteristics of the seismic data spectrum. The optimal threshold curve to adapt to the different seismic data is obtained by the variable parameters to further improve the calculation accuracy and computational efficiency of the algorithm. In order to achieve adaptive attenuation of random noise during reconstruction and to improve signal-to-noise ratio of reconstructed seismic data, a data-driven weighted regression coefficient calculation method is proposed. By using the percentage of the data-driven threshold to the threshold interval in each iteration Weighted back-up coefficient. The new method is applied to simulate three-dimensional data and actual three-dimensional seismic data. The analysis results show that inverse proportion thresholding has obvious advantages over traditional threshold in improving the data reconstruction efficiency and accuracy. The newly proposed weighted additive coefficient adding strategy can be effective Improve the signal to noise ratio of reconstructed data.