一种可靠的强噪声三分量微地震数据初至拾取方法

精确的初至拾取是微地震数据处理中至关重要的环节。主流的长短时窗比法(STA/LTA)和基于自回归模型的赤池信息准则(AR-AIC)方法,对强噪声数据的拾取效果并不理想。为了更为精确的估计强噪声数据初至,提出了一种基于小波多尺度分析(WMA)和AIC算法的联合拾取方法。首先使用WMA对强噪声三分量(3C)微地震数据进行分解,并重构其近似数据作为实际计算数据,同时计算其绝对值的最大值点,来约束AIC计算数据段,最终选取AIC序列的全局最小值点作为其初至点。文中采用合成数据和实测数据对该改进算法进行了验证,拾取结果表明该算法能有效适用于强噪声微地震数据初至拾取,并明显提高其拾取精度(误差在0.25～0.5 ms之间)。

A reliable first arrival picking method for high-noise three component microseismic data

Accurate first-arrival picking plays a significant role in microseismic data processing.The dominate methods,i.e.,short-term average to long-term average ratio( STA/LTA) and the autoregressive( AR) model using Akaike Information Criterion( AIC) algorithm,are not optimal for picking up high-noise data.In order to optimize the first-arrival point estimation of high-noise data,the authors propose an improved picking method based on wavelet multi-scale analysis( WMA) and AIC algorithm.In this algorithm,WMA is used to decompose the high-noise three-component( 3 C) microseismic data,and the actual calculation data are reconstructed based on the approximation data.Then the maximum value of the absolute value is calculated to constrain the AIC calculation data segment;on such a basis,the global minimum value of the AIC sequence is selected as its first arrival point.The improved algorithm is verified by both synthetic data and field data in this paper.The results show that the improved first-arrival picking algorithm can be effectively applied to high-noise three component microseismic data processing and greatly enhance its accuracy( the error range is 0.25 ~ 0.5 ms).