One of the most important goals of seismic stratigraphy studies is to interpret the elements of the seismic facies with respect to the geologic environment. Prestack seismic data carry rich information that can help us get higher resolution and more accurate facies maps. Therefore, it is promising to use prestack seismic data for the seismic facies recognition task. However, because each identified object changes from the poststack trace vectors to a prestack trace matrix, effective feature extraction becomes more challenging. We have developed a novel data-driven offset-temporal feature extraction approach using the deep convolutional autoencoder (DCAE). As an unsupervised deep learning method, DCAE learns nonlinear, discriminant, and invariant features from unlabeled data. Then, seismic facies analysis can be accomplished through the use of conventional classification or clustering techniques (e.g., K-means or self-organizing maps). Using a physical model and field prestack seismic surveys, we comprehensively determine the effectiveness of our scheme. Our results indicate that DCAE provides a much higher resolution than the conventional methods and offers the potential to significantly highlight stratigraphic and depositional information.

Deconvolution is an essential step in seismic data processing. Sparse-spike deconvolution often is used to enhance the resolution of the seismic image by adding a model-driven regularization term. However, this method does not consider the features of the data, nor does it exactly describe the relationship between seismic data and the desired attribute (such as seismic reflectivity). We have developed a data and model dual-driven seismic deconvolution method based on error-constrained joint sparse representation (SR) using borehole measurement and surface seismic data. The combined features of the borehole reflectivity and the surface seismic data can be obtained through joint dictionary learning. With the help of the joint dictionary, the relationship between seismic waveforms and reflectivity is captured by the sparse coefficients. We construct the regularization term of deconvolution by alternately decomposing the error of the synthesized data via sparse reconstruction and the observed seismic data. Unlike model-driven methods, the constraint term of the new method can be established by the error-constrained SR. Based on this SR, the initial model of reflectivity is obtained to realize the sparse deconvolution of seismic data under the constraint of borehole data features. In general, this method is a data and model dual-driven deconvolution. Synthetic and field data tests demonstrate that this method can effectively improve the resolution and accuracy of deconvolution.

We have addressed the seismic impedance inversion problem, which is often ill posed because of inaccurate and insufficient data. The approach taken is based on dictionary learning and sparse representation. By shifting a patch window of fixed size on all well-log data, a large number of small overlapping patches are generated. Then regarding these patches as a training set and using K-singular value decomposition algorithm, we obtain a dictionary that describes the common features of subsurface models within the current survey area. On the basis of the assumption that the subsurface geology has similarity and lateral continuity to some extent, the dictionary is used to approximate each model via sparse representations over the learned dictionary. In particular, we impose the sparse representations as additional constraints into the inversion procedure, leading to a well-defined objective function that can not only fit the observed seismic data but also honor the features of the well-log data. We adopt a coordinate descent strategy to solve this objective function. Meanwhile, to enforce lateral continuity in the inverted models, we use an additional stage in which we use the nonlocal similarity information that is extracted from seismic data as spatial coherent prior knowledge to refine the estimated models. Compared with several traditional impedance inversion methods, our algorithm can produce solutions of much higher quality qualitatively and quantitatively.

It is crucial to retain as many signal components as possible, while noise is eliminated for wavelet packet noise reduction algorithms. Conventional thresholding functions take coefficients smaller than a given threshold as the noise component and get them removed in many ways. In this letter, considering the potential probability of being the signal-related component of coefficients whose value is smaller than the given threshold and the nonnegligible fact that microseismic (MS) event is sparse compared with the noise, we proposed a new wavelet packet-based denoising method via fuzzy partition. First, instead of a hard partition from a given threshold, wavelet packet coefficients get reduced by a fuzzy partition to retain more potential signal elements and suppress the noise. Second, we employ fuzzy c-means (FCM) clustering to identify the interval period of the MS event further to remove the residual noise and additional resonance in processed time series. We tested our method on synthetic datasets and real-field data from an MS monitoring experiment in a coal mine in Sichuan Basin, China. We utilize Pearson correlation coefficient and root-mean-square error between ideal signal and denoised data as performance indicators in synthetic tests, while sample entropy and kurtosis of denoised data are involved in the real field dataset. Test results from synthetic datasets and the real field dataset demonstrate that the proposed noise reduction method is superior to traditional hard-, soft-, and garrote-thresholding and is more applicable and effective in MS data processing.