最优小波包变换在地震信号去噪中的应用

应用最优小波包变换,采用取决于节点噪声时频特征的相关阈值,对模拟地震波到达的加噪合成信号、SDAES数字声发射仪采集的微震实验信号、NIED观测台站记录的日本2007年能登半岛地震信号进行去噪,并和基于小波变换的其他方法进行了去噪效果比较.结果表明,该方法获得的信号信噪比(SNR)高、失真低,体现出了总体优越性.     

基于压缩感知的小波阈值和CEEMD联合去噪方法

中深层地质条件复杂,地震资料品质差,主要表现为:地震资料信噪比低、有效信号弱.如何在去噪的同时有效保留弱有效信号,获取高信噪比的地震数据成为地震数据处理的关键问题.传统小波阈值与互补集合经验模态分解(CEEMD)联合去噪方法相比单一方法可以获取更高品质的地震数据.基于压缩感知理论的去噪方法利用地震数据在变换域中的稀疏特性,通过设定稀疏基矩阵和测量矩阵,可以将地震数据去噪问题转化成求解最优化问题,通过最优解重构原始信号,实现对地震资料的去噪处理.该方法能够在有效衰减随机噪声的同时最大限度的保留有效信号.本文基于压缩感知理论开展小波阈值去噪方法研究,并在此基础上结合CEEMD方法对含噪较多的固有模态分量进行有针对性的随机噪声压制.通过对含噪数据开展不同方法的去噪结果对比可见,本文方法可以在保证高信噪比的基础上更为有效的保留弱有效信号,数值试算验证了该方法对弱有效信号地震数据去噪具有显著优势.     

基于自适应Morlet小波的低信噪比微震信号初至拾取方法研究

针对影响微震初至拾取的资料信噪比过低的问题,传统方法的拾取精度与稳定性大多不太理想.为了克服低信噪比条件下初至无法有效拾取的缺点,本文设计了一种对目标成分具有高敏感性的自适应Morlet小波基,通过利用该小波基对微震记录进行小波分解,利用三分量数据的有效成分在小波域内具有特征相关性,对三分量小波系数进行主成分分析,提取主成分特征,最终对各级主成分进行加权重构,实现对低信噪比微震信号的初至拾取.在设计有不同信噪比的模型实验与实际资料应用中,该算法均表现出优异的抗噪性能.在极低信噪比条件下,仍能精确指示有效成分的初至.模型实验与实际资料处理结果均验证了本方法对极低信噪比微震资料的初至拾取处理上的有效性与实用性.本方法在微震监测等相关领域中具有较高的理论与应用价值.     

基于提升算法和百分位数软阈值的小波去噪技术

在地震勘探领域,随机噪声一直是影响地震信号信噪比的主要因素之一,如何从被干扰的地震信号中有效去除随机噪声并保护有用信号具有重要的意义.针对经典小波变换在计算效率方面的缺陷,本文推荐应用提升算法实现第二代小波变换的构建,分析和对比了提升算法(Lifting Scheme)下不同小波变换方法的特性,选取更加符合小波域去噪原理的CDF 9/7双正交小波变换作为基本算法,同时应用了简单、有效的百分位数(Percentiles)软阈值进行信噪分离.通过理论模型处理,本方法可以在去噪能力和保护有用信号之间找到很好的平衡点.实际剖面的处理效果表明,此方法不仅能有效的滤除随机噪声,而且很好地保护有用信号,提高地震数据分析的精确性.     

基于Curvelet变换的地震资料信噪分离技术

在地震资料中,噪声干扰严重影响了有效信号的提取,为此必须进行信噪分离处理.本文提出一种基于Curvelet变换和KL变换相结合的软硬阈值折衷处理方法.首先对地震数据进行Curvelet变换,然后对各尺度系数选取适当阈值压制噪声干扰,再利用KL变换提取数据中的相干有效信号,最后重构得到去噪后的记录.经合成记录和实际地震资料处理实验证明,该方法与小波变换法相比较,更能有效进行信噪分离,提高地震剖面信噪比和分辨率.     

基于地震信号波形形态差异的面波噪声稀疏优化分离方法

实际地震信号通常可表示为具有波形特征差异的多种基本波形信号的线性组合,如叠前道集中的工频干扰噪声与有效波信号、面波噪声与体波信号等.选择单一数学变换方法,往往不易实现地震信号的稀疏表示.近年来发展的形态成分分析理论,通过联合多种数学变换,可实现对复杂信号的稀疏表示.本文根据单道地震记录中面波与体波信号波形结构特征的差异性,提出一种基于形态成分分析的面波噪声衰减方法.针对面波的低频、窄带以及频散特性选择一维平稳小波变换作为其稀疏表示字典,而针对体波波形的局部相关特性选择局部离散余弦变换作为其稀疏表示字典,建立基于双波形字典的形态成分分析模型,通过求解该稀疏优化问题获得最终的信噪分离结果.理论模型和实际地震资料处理证实该方法不仅能够衰减单炮地震记录中的强面波干扰噪声,同时能够更好地保护有效信号的波形特征与频谱带宽,为地震资料的后续处理和分析提供良好的数据基础.     

基于分数阶小波域GSM模型的地震信号随机噪声压制方法

地震信号中的随机噪声是一种干扰波,严重降低了地震信号的信噪比,并影响着资料的后续处理和分析.本文根据地震信号中有效信号和随机噪声的差异,结合分数阶B样条小波变换与高斯尺度混合模型提出了一种地震信号随机噪声压制方法.首先利用分数阶B样条小波变换将含噪地震信号映射到最优分数阶小波时频域内,然后对各小波子带系数分别建立高斯尺度混合模型,由贝叶斯方法估计出源地震信号小波系数,最后使用分数阶B样条小波逆变换重构得到降噪后的地震信号.利用本文方法对合成地震记录和实际地震信号进行降噪处理,实验结果表明本文方法能够有效地压制地震信号中的随机噪声,并且较好地保留了有效信号.     

基于Seislet-TV双正则化约束的地震随机噪声压制方法

人工地震数据总是受到随机噪声的干扰,地震数据时-空变的特性使得常规去噪方法处理效果并不理想,容易导致有效信号的损失.目前广泛应用的预测滤波类方法存在处理时变数据能力不足的问题.随着压缩感知理论的不断完善,稀疏变换阈值算法能够解决时变地震数据噪声压制问题,但是常规的稀疏变换方法,如傅里叶变换,小波变换等,并不是特殊针对地震数据设计的,很难提供地震数据最佳的压缩特征,同时,常规阈值算法容易导致去噪结果过于平滑.因此开发更加有效的时-空变地震数据信噪分离方法具有重要的工业价值.本文将地震数据信噪分离问题归纳为数学基追踪问题,在压缩感知理论框架下,利用特殊针对地震数据设计的VD-seislet稀疏变换方法,结合全变差(TV)算法,构建seislet-TV双正则化条件,并利用分裂Bregman迭代算法求解约束最优化问题,实现地震数据的有效信噪分离.通过理论模型和实际数据测试本文方法,并且与工业标准FXdecon方法进行比较,结果表明基于seislet-TV双正则化约束条件的迭代方法能够更加有效地保护时-空变地震信号,压制地震数据中的强随机噪声.     

多小波方法及其在深地震测深数据去噪中的应用

在宽角反射/折射地震测深数据处理中,仍多用基于傅里叶变换的滤波方法和小波去噪方法。鉴于傅里叶方法对稳态信号很有效但对非稳态的地震信号效果不佳的状况以及小波不能同时具有正交性、紧支性、对称性,本文给出了基于多小波的去噪方法,多小波具有正交性、对称性、紧支性,克服了传统小波的缺陷。编写了多小波去噪方法的人机交互软件。该软件可以方便快捷地显示宽角反射/折射地震记录截面,进行多小波域的阈值去噪。实例计算结果表明,本文所述方法和编写的软件有效且可行。     

基于广义Beta小波稀疏域混合约束优化的地震去噪方法研究

地震随机噪声压制是鄂尔多斯盆地黄土塬、沙漠、戈壁滩等复杂地表区域低信噪比地震资料处理的一项重要任务.稀疏反演去噪是地震随机噪声压制的常用方法之一.?₁范数和全变分(Total Variation, TV)正则化是稀疏变换域去噪方法中常用的两种正则化项.但是,?₁范数是对?₀范数的松弛,难以提供更稀疏的去噪结果;基于TV正则化项的方法容易引起阶梯状异常结果.因此,为了避免上述缺点,本文提出了一种基于广义Beta小波稀疏域混合范数优化的地震随机噪声压制方法和算法流程实现.首先利用广义Beta小波紧标架加快计算,获得具有更高局域化性的稀疏时频表示.其次是引入包括?_p范数和TV正则化的混合约束项,克服单一正则化项的缺点.最后,利用鄂尔多斯盆地黄土塬区的合成地震数据、三维叠后地震数据和共反射点道集数据验证了本文去噪方法的有效性.结果表明：本文提出的去噪方法既能够有效抑制随机噪声、显著提高信噪比,让地震同相轴连续光滑;又能够准确保护有效信号,保持波组间的相对幅值,突出有利微小断层和含油气层的振幅形态.     

基于Context模型的Shearlet变换地面微地震随机噪声压制

地面微地震监测采集到的微地震信号通常能量微弱,信噪比低,如何提高微震数据的信噪比是数据处理的难题.Shearlet变换是一种新型的多尺度几何分析方法,具有敏感的方向性和较强的稀疏表示特性,能起到很好的随机噪声压制效果.由于地面微震数据的有效信号大多被淹没在噪声中,基于传统阈值的Shearlet变换（the traditional threshold-based Shearlet transform TST）只考虑到尺度或方向的阈值,在去噪过程中会过度扼制有效信号系数,造成有效信号能量损失.因而,本文建立Context模型,得到基于Context模型的Shearlet变换（the Context-model-based Shearlet transform CMST）方法,改进传统Shearlet阈值方法的不足.我们通过所建立的Context模型将能量相近的各方向系数划分为同一组,并分组估计阈值,分别处理各部分系数,达到微弱同相轴有效恢复的目的.通过TST及CMST的模拟实验与实际地面微震记录处理结果对比可知,本文方法在低信噪比条件下比对比方法更加有效地恢复地面微震数据的微弱信号,随机噪声压制效果明显,在-10 dB条件下,提升信噪比18.3741 dB.     

Full‐waveform based microseismic event detection and signal enhancement: An application of the subspace approach

Microseismic monitoring has proven invaluable for optimizing hydraulic fracturing stimulations and monitoring reservoir changes. The signal to noise ratio of the recorded microseismic data varies enormously from one dataset to another, and it can often be very low, especially for surface monitoring scenarios. Moreover, the data are often contaminated by correlated noises such as borehole waves in the downhole monitoring case. These issues pose a significant challenge for microseismic event detection. In addition, for downhole monitoring, the location of microseismic events relies on the accurate polarization analysis of the often weak P‐wave to determine the event azimuth. Therefore, enhancing the microseismic signal, especially the low signal to noise ratio P‐wave data, has become an important task. In this study, a statistical approach based on the binary hypothesis test is developed to detect the weak events embedded in high noise. The method constructs a vector space, known as the signal subspace, from previously detected events to represent similar, yet significantly variable microseismic signals from specific source regions. Empirical procedures are presented for building the signal subspace from clusters of events. The distribution of the detection statistics is analysed to determine the parameters of the subspace detector including the signal subspace dimension and detection threshold. The effect of correlated noise is corrected in the statistical analysis. The subspace design and detection approach is illustrated on a dual‐array hydrofracture monitoring dataset. The comparison between the subspace approach, array correlation method, and array short‐time average/long‐time average detector is performed on the data from the far monitoring well. It is shown that, at the same expected false alarm rate, the subspace detector gives fewer false alarms than the array short‐time average/long‐time average detector and more event detections than the array correlation detector. The additionally detected events from the subspace detector are further validated using the data from the nearby monitoring well. The comparison demonstrates the potential benefit of using the subspace approach to improve the microseismic viewing distance. Following event detection, a novel method based on subspace projection is proposed to enhance weak microseismic signals. Examples on field data are presented, indicating the effectiveness of this subspace‐projection‐based signal enhancement procedure.     

A first arrival detection method for low SNR microseismic signal

Most of the microseismic signals have low signal-to-noise ratio (SNR) due to the strong background noise, which makes it difficult to locate the first arrival time. Both accuracy and stability of conventional methods are poor in this situation. To overcome this problem, here we proposed a new method based on the adaptive Morlet wavelet and principal component analysis process in wavelet coefficients matrix. The three components of microseismic signal make it possible to extract the features in wavelet coefficients domain. Then the reconstructed signal from weighted features presents an obvious first arrival. Tests on synthetic signals and real data provide a solid evidence for its feasibility in low SNR microseismic signal.     

基于小波变换的大尺度岩体结构微震监测信号去噪方法研究

为将小波去噪方法应用于大尺度岩体结构微震监测信号的去噪研究,首先在MATLAB环境下进行仿真,验证了使用Symlet6小波进行小波去噪的可行性;利用4种自适应阈值规则对含噪信号进行去噪对比,结果表明4种阈值去噪后的信号在均方差较小的情况下都极大地提高了信号的信噪比,有效地去除了噪声,对不同的含噪信号,无偏似然原则阈值去...     

Consistency of stress state,locations and source mechanisms of events induced by hydraulic fracturing: downhole monitoring

We present results of processed microseismic events induced by hydraulic fracturing and detected using dual downhole monitoring arrays. The results provide valuable insight into hydraulic fracturing. For our study, we detected and located microseismic events and determined their magnitudes, source mechanisms and inverted stress field orientation. Event locations formed a distinct linear trend above the stimulated intervals. Source mechanisms were only computed for high‐quality events detected on a sufficient number of receivers. All the detected source mechanisms were dip‐slip mechanisms with steep and nearly horizontal nodal planes. The source mechanisms represented shear events and the non‐double‐couple components were very small. Such small, non‐double‐couple components are consistent with a noise level in the data and velocity model uncertainties. Strikes of inverted mechanisms corresponding to the nearly vertical fault plane are (within the error of measurements) identical with the strike of the location trend. Ambient principal stress directions were inverted from the source mechanisms. The least principal stress, σ₃, was determined perpendicular to the strike of the trend of the locations, indicating that the hydraulic fracture propagated in the direction of maximum horizontal stress. Our analysis indicated that the source mechanisms observed using downhole instruments are consistent with the source mechanisms observed in microseismic monitoring arrays in other locations. Furthermore, the orientation of the inverted principal components of the ambient stress field is in agreement with the orientation of the known regional stress, implying that microseismic events induced by hydraulic fracturing are controlled by the regional stress field.     

基于时频稀疏性分析法的低信噪比微震事件识别与恢复

微震监测是直观评价压裂过程和压裂效果的有效手段.微震事件识别是微震监测的首要步骤.然而对于低信噪比微震监测数据,常规识别方法很难取得满意效果.基于微震事件在时频域中的稀疏性,本文提出利用Renyi熵值表示微震监测数据的时频稀疏程度,并以时频距离为约束条件,建立以低熵值的道数为判别阈值的目标函数.本文方法能在识别出微震事件的同时,恢复出较为清晰的微震事件.通过数值计算和对实际监测数据的测试,表明该方法对低信噪比的微震监测数据有较好的处理效果.     

Automatic event detection in low SNR microseismic signals based on multi-scale permutation entropy and a support vector machine

Microseismic monitoring is an effective means for providing early warning of rock or coal dynamical disasters, and its first step is microseismic event detection, although low SNR microseismic signals often cannot effectively be detected by routine methods. To solve this problem, this paper presents permutation entropy and a support vector machine to detect low SNR microseismic events. First, an extraction method of signal features based on multi-scale permutation entropy is proposed by studying the influence of the scale factor on the signal permutation entropy. Second, the detection model of low SNR microseismic events based on the least squares support vector machine is built by performing a multi-scale permutation entropy calculation for the collected vibration signals, constructing a feature vector set of signals. Finally, a comparative analysis of the microseismic events and noise signals in the experiment proves that the different characteristics of the two can be fully expressed by using multi-scale permutation entropy. The detection model of microseismic events combined with the support vector machine, which has the features of high classification accuracy and fast real-time algorithms, can meet the requirements of online, real-time extractions of microseismic events.     

Microseismic event denoising via adaptive directional vector median filters

We present a novel denoising scheme via Radon transform-based adaptive vector directional median filters named adaptive directional vector median filter (AD-VMF) to suppress noise for microseismic downhole dataset. AD-VMF contains three major steps for microseismic downhole data processing: (i) applying Radon transform on the microseismic data to obtain the parameters of the waves, (ii) performing S-transform to determine the parameters for filters, and (iii) applying the parameters for vector median filter (VMF) to denoise the data. The steps (i) and (ii) can realize the automatic direction detection. The proposed algorithm is tested with synthetic and field datasets that were recorded with a vertical array of receivers. The P-wave and S-wave direct arrivals are properly denoised for poor signal-to-noise ratio (SNR) records. In the simulation case, we also evaluate the performance with mean square error (MSE) in terms of signal-to-noise ratio (SNR). The result shows that the distortion of the proposed method is very low; the SNR is even less than 0 dB.     

Weak Seismic Signal Extraction Based on the Curvelet Transform

Seismic signal denoising is a key step in seismic data processing. Airgun signals are easy to be interfered with by noise when it travels a long distance due to the weak energy of active source signal of the airgun. Aiming to solve this problem, and considering that the conventional Curvelet transform threshold processing method does not use the seismic spectrum information, we independently process the Curvelet scale layer corresponding to valid data based on the characteristics of the Curvelet transform of multi-scale, multi-direction and capable of expressing the sparse seismic signals in order to fully excavate the information features. Combined with the Curvelet adaptive threshold denoising the algorithm, we apply the Curvelet transform to denoising seismic signals while retaining the weak information in the signal as much as possible. The simulation experiments show that the improved threshold denoising method based on Curvelet transform is superior to the frequency domain filtering, wavelet denoising and traditional Curvelet denoising method in detailed information extraction and signal denoising of low SNR signals. The calculation accuracy of the relative wave velocity variation of underground medium is improved.     

Identification of blasting vibration and coal-rock fracturing microseismic signals

Α new method based on variational mode decomposition (VMD) is proposed to distinguish between coal-rock fracturing and blasting vibration microseismic signals. First, the signals are decomposed to obtain the variational mode components, which are ranked by frequency in descending order. Second, each mode component is extracted to form the eigenvector of the energy of the original signal and calculate the center of gravity coefficient of the energy distribution plane. Finally, the coal-rock fracturing and blasting vibration signals are classified using a decision tree stump. Experimental results suggest that VMD can effectively separate the signal components into coal-rock fracturing and blasting vibration signals based on frequency. The contrast in the energy distribution center coefficient after the dimension reduction of the energy distribution eigenvector accurately identifies the two types of microseismic signals. The method is verified by comparing it to EMD and wavelet packet decomposition.