Layer | Fill | Outline |
---|
Map layers
Theme | Visible | Selectable | Appearance | Zoom Range (now: 0) |
---|
Fill | Stroke |
---|---|
Collaborating Authors
Bestagini, Paolo
An alternative approach has been presented by Ulyanov et al. (2018), showing that a CNN can learn the inner structure of a Interpolation of seismic data is an important pre-processing 2D image from the corrupted image itself, with no additional step in most seismic processing workflows. Through the deep pre-training, in order to perform tasks such as inpainting or image prior paradigm, it is possible to use Convolutional Neural super-resolution. Through this approach it is possible to exploit Networks for seismic data interpolation without the costly the features of the specific CNN architecture acting as a and prone-to-overfitting training stage. The proposed method prior for the inverse problem, thus avoiding many of the overfitting makes use of the multi-res U-net architecture as a deep prior to issues due to insufficient training strategies.
Wavefield compression for seismic imaging via convolutional neural networks
Devoti, Francesco (Politecnico di Milano, Italy) | Parera, Claudia (Politecnico di Milano, Italy) | Lieto, Alessandro (Politecnico di Milano, Italy) | Moro, Daniele (Politecnico di Milano, Italy) | Lipari, Vincenzo (Politecnico di Milano, Italy) | Bestagini, Paolo (Politecnico di Milano, Italy) | Tubaro, Stefano (Politecnico di Milano, Italy)
ABSTRACT FullWaveform Inversion and Reverse Time Migration are usually based on the adjoint state method and rely on the ready availability of wavefield snapshots. Therefore, standard algorithms store the full source wavefield on disk. This makes Full Waveform Inversion and Reverse Time Migration techniques particularly demanding in terms of disk input/output performance. To face this issue, a common solution is to compress wavefield information in order to reduce input/output operations overhead. In this paper we propose a couple of wavefield compression methods based on Convolutional Neural Networks (CNNs). Specifically, a convolutional autoencoder is trained to compress wavefield snapshots and a specifically designed U-net is trained to reconstruct (i.e. interpolate) the wavefield in the temporal dimension. Results show that these promising techniques could help decreasing storage needs for wavefield snapshots. This makes it possible to balance the required signal-to-noise ratio and compression gain. Presentation Date: Monday, September 16, 2019 Session Start Time: 1:50 PM Presentation Time: 3:55 PM Location: 221D Presentation Type: Oral
- Research Report > New Finding (0.34)
- Research Report > Experimental Study (0.34)
Abstract The advent of new deep-learning and machine-learning paradigms enables the development of new solutions to tackle the challenges posed by new geophysical imaging applications. For this reason, convolutional neural networks (CNNs) have been deeply investigated as novel tools for seismic image processing. In particular, we have studied a specific CNN architecture, the generative adversarial network (GAN), through which we process seismic migrated images to obtain different kinds of output depending on the application target defined during training. We have developed two proof-of-concept applications. In the first application, a GAN is trained to turn a low-quality migrated image into a high-quality one, as if the acquisition geometry was much more dense than in the input. In the second example, the GAN is trained to turn a migrated image into the respective deconvolved reflectivity image. The effectiveness of the investigated approach is validated by means of tests performed on synthetic examples.
- Reservoir Description and Dynamics > Reservoir Characterization > Seismic processing and interpretation (1.00)
- Data Science & Engineering Analytics > Information Management and Systems > Neural networks (1.00)
- Data Science & Engineering Analytics > Information Management and Systems > Artificial intelligence (1.00)
Seismic data interpolation through convolutional autoencoder
Mandelli, Sara (Politecnico di Milano, Italy) | Borra, Federico (Politecnico di Milano, Italy) | Lipari, Vincenzo (Politecnico di Milano, Italy) | Bestagini, Paolo (Politecnico di Milano, Italy) | Sarti, Augusto (Politecnico di Milano, Italy) | Tubaro, Stefano (Politecnico di Milano, Italy)
ABSTRACT A common issue of seismic data analysis consists in the lack of regular and densely sampled seismic traces. This problem is commonly tackled by rank optimization or statistical features learning algorithms, which allow interpolation and denoising of corrupted data. In this paper, we propose a completely novel approach for reconstructing missing traces of pre-stack seismic data, taking inspiration from computer vision and image processing latest developments. More specifically, we exploit a specific kind of convolutional neural networks known as convolutional autoencoder. We illustrate the advantages of using deep learning strategies with respect to state-of-the-art by comparing the achieved results over a well-known seismic dataset. Presentation Date: Wednesday, October 17, 2018 Start Time: 1:50:00 PM Location: 204C (Anaheim Convention Center) Presentation Type: Oral
ABSTRACT The new challenges of geophysical imaging applications ask for new methodologies going beyond the standard and well established techniques. In this work we propose a novel tool for seismic imaging applications based on recent advances in deep neural networks. Specifically, we use a generative adversarial network (GAN) to process seismic migrated images in order to potentially obtain different kinds of outputs depending on the application target at training stage. We demonstrate the promising features of this tool through a couple of synthetic examples. In the first example, the GAN is trained to turn a low-quality migrated image into a high-quality one, as if the acquisition geometry were much more dense than in the input. In the second example, the GAN is trained to turn a migrated image into the respective deconvolved reflectivity image. Presentation Date: Wednesday, October 17, 2018 Start Time: 9:20:00 AM Location: Poster Station 1 Presentation Type: Poster
ABSTRACT In this work we describe a machine learning pipeline for facies classification based on wireline logging measurements. The algorithm has been designed to work even with a relatively small training set and amount of features. The method is based on a gradient boosting classifier which demonstrated to be effective in such a circumstance. A key aspect of the algorithm is feature augmentation, which resulted in a significant boost in accuracy. The algorithm has been tested also through participation to the SEG machine learning contest. Presentation Date: Wednesday, September 27, 2017 Start Time: 3:05 PM Location: 350D Presentation Type: ORAL
- Geology > Rock Type > Sedimentary Rock > Carbonate Rock (0.69)
- Geology > Rock Type > Sedimentary Rock > Clastic Rock (0.48)