Semi-supervised impedance inversion by Bayesian neural network based on 2-d CNN pre-training

Ge, Muyang; Wang, Wenlong; Zheng, Wangxiangming

doi:10.1190/iwmg2021-34.1

Cited by 4 publications

(2 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The method depends on having enough labels to ensure the stability of the Generative Adversarial Networks (GAN) [29], so the inversion of the SEAM Phase I (Fig. 4, 7) still requires 34 logging labels, and its predictions have significant discontinuities in the horizontal direction, the same drawback is also reflected in other semi-supervised impedance inversion methods [30]- [32]. The discontinuity in the horizontal direction is due to the fact that these methods try to match the dimensionality of the logs by downscaling the 3D or 2D seismic data, which is avoided by the multidimensional inversion proposed by Wu et al [11], the idea of this method originates from medical image segmentation [33], where the model is trained using labels of the same dimensionality as the seismic and the weights of the unlabeled regions are set to zero.…”

Section: Introductionmentioning

confidence: 99%

ContrasInver: Voxel-wise Contrastive Semi-supervised Learning for Seismic Inversion

Dou¹,

Li²,

Li³

et al. 2023

Preprint

View full text Add to dashboard Cite

Recent studies have shown that learning theories have been very successful in hydrocarbon exploration. Inversion of seismic into various attributes through the relationship of 1D well-logs and 3D seismic is an essential step in reservoir description, among which, acoustic impedance is one of the most critical attributes, and although current deep learningbased impedance inversion obtains promising results, it relies on a large number of logs (1D labels, typically more than 30 well-logs are required per inversion), which is unacceptable in many practical explorations. In this work, we define acoustic impedance inversion as a regression task for learning sparse 1D labels from 3D volume data and propose a voxel-wise semisupervised contrastive learning framework, ContrasInver, for regression tasks under sparse labels. ConstraInver consists of several key components, including a novel pre-training method for 3D seismic data inversion, a contrastive semi-supervised strategy for diffusing well-log information to the global, and a continuous-value vectorized characterization method for a contrastive learning-based regression task, and also designed the distance TopK sampling method for improving the training efficiency. We performed a complete ablation study on SEAM Phase I synthetic data to verify the effectiveness of each component and compared our approach with the current mainstream methods on this data, and our approach demonstrated very significant advantages. In this data we achieved an SSIM of 0.92 and an MSE of 0.079 with only four well-logs. ConstraInver is the first purely data-driven approach to invert two classic field data, F3 Netherlands (only four well-logs) and Delft (only three well-logs) and achieves very reasonable and reliable results.

show abstract

Section: Introductionmentioning

confidence: 99%

ContrasInver: Voxel-wise Contrastive Semi-supervised Learning for Seismic Inversion

Dou¹,

Li²,

Li³

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Many 1D networks are also incapable of capturing sudden changes in the rock properties in the complex geological structure due to the limited training well logs [21]. The increasing number of publications [4,[22][23] and the industry are keeping an eye on the 2D network implementations for seismic impedance inversion which shows a positive trend in this field. In this paper, we use two-dimensional (2D) convolution for training, which enhances the continuity with the help of stratum structure correlation.…”

Section: Introductionmentioning

confidence: 99%

Attention and Hybrid Loss Guided 2-D Network for Seismic Impedance Inversion

Xie

Ye³

2023

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Deep learning (DL) methods, especially Convolutional Neural Networks (CNNs), achieve state-of-the-art performance on seismic impedance inversion. Most of the methods are based on one-dimensional (1D) convolution, tending to yield lateral discontinuities of impedance on field data applications. To alleviate this problem, we design a network equipped with two-dimensional (2D) convolutions and a coordinate attention (CA) block. The former can take the relationship between adjacent traces into consideration. The latter can capture the positional relationship of the geological structure, both horizontally and vertically. At the same time, we use a hybrid loss combined with edge operator and Mean Square Error (MSE) to further improve the stability of the designed network. Comparison experiments on synthetic SEAM model and field seismic data demonstrate the effectiveness of the adopted components, 2D convolution, CA and hybrid loss function, in improving the lateral continuity of inverted impedance. For field seismic data, the impedance predicted by proposed method shows improved lateral continuity and high resolution compared with the 1D network and Constrained Sparse Spike Inversion (CSSI) method using commercial software (InverTrace Plus module in Jason).

show abstract

Pre-stack seismic inversion based on one-dimensional GRU combined with two-dimensional improved ASPP

Chen,

Li,

Wei

et al. 2024

Journal of Geophysics and Engineering

View full text Add to dashboard Cite

Pre-stack seismic inversion is essential to detailed stratigraphic interpretation of seismic data. Recently, various deep learning methods have been introduced into pre-stack inversion, effectively capturing the vertical correlations of seismic data. However, existing deep learning methods face challenges such as insufficient feature extraction, poor lateral continuity, and unclear inversion details. We introduce the atrous spatial pyramid pooling (ASPP) module into the pre-stack inversion process, modifying the connection order and mode of its three components. Additionally, we incorporate a triplet attention module to extract features at different scales and utilize a gate recurrent unit (GRU) module to extract global information. During the network training stage, we employ a multi-gather simultaneous inversion method, combining one-dimensional and two-dimensional inversions. The proposed method is named as IGIT (I for Improved ASPP, G for GRU, I for Initial model and T for Triplet attention). To verify the feasibility of this network model, we evaluate it using the Marmousi2 model, SEAM model, and field data, comparing the results with other deep learning methods. Experimental results demonstrate that the IGIT not only improves lateral continuity but also delivers accurate and clear inversion details. Notably, the inversion effect for density parameters shows significant enhancement.

show abstract

Semi-supervised impedance inversion by Bayesian neural network based on 2-d CNN pre-training

Cited by 4 publications

References 6 publications

ContrasInver: Voxel-wise Contrastive Semi-supervised Learning for Seismic Inversion

ContrasInver: Voxel-wise Contrastive Semi-supervised Learning for Seismic Inversion

Attention and Hybrid Loss Guided 2-D Network for Seismic Impedance Inversion

Pre-stack seismic inversion based on one-dimensional GRU combined with two-dimensional improved ASPP

Contact Info

Product

Resources

About