ALBERT

Description: 〈span〉〈div〉Abstract〈/div〉Detecting a specific horizon in seismic images is a valuable tool for geologic interpretation. Because hand picking the locations of the horizon is a time-consuming process, automated computational methods were developed starting three decades ago. Until now, most networks have been trained on data that were created by cutting larger seismic images into many small patches. This limits the networks ability to learn from large-scale geologic structures. Moreover, currently available networks and training strategies require label patches that have full and continuous horizon picks (annotations), which are also time-consuming to generate. We have developed a projected loss function that enables training on labels with just a few annotated pixels and has no issue with the other unknown label pixels. We use this loss function for training convolutional networks with a multiresolution structure, including variants of the U-net. Our networks learn from a small number of large seismic images without creating patches. Training uses all seismic data without reserving some for validation. Only the labels are split into training/testing. We validate the accuracy of the trained network using the horizon picks that were never shown to the network. Contrary to other work on horizon tracking, we train the network to perform nonlinear regression, not classification. As such, we generate labels as the convolution of a Gaussian kernel and the known horizon locations that communicate uncertainty in the labels. The network output is the probability of the horizon location. We examine the new method on two different data sets, one for horizon extrapolation and another data set for interpolation. We found that the predictions of our methodology are accurate even in areas far from known horizon locations because our learning strategy exploits all data in large seismic images.〈/span〉

Description: 〈span〉〈div〉Abstract〈/div〉There has been a surge of interest in neural networks for the interpretation of seismic images over the last few years. Network-based learning methods can provide fast and accurate automatic interpretation, provided that there are many training labels. We provide an introduction to the field for geophysicists who are familiar with the framework of forward modeling and inversion. We explain the similarities and differences between deep networks and other geophysical inverse problems and show their utility in solving problems such as lithology interpolation between wells, horizon tracking, and segmentation of seismic images. The benefits of our approach are demonstrated on field data from the Sea of Ireland and the North Sea.〈/span〉

Description: Full-waveform inversion is challenging in complex geologic areas. Even when provided with an accurate starting model, the inversion algorithms often struggle to update the velocity model. Compared with other areas in applied geophysics, including prior information in full-waveform inversion is still in its relative infancy. In part, this is due to the fact that it is difficult to incorporate prior information that relates to geologic settings where strong discontinuities in the velocity model dominate, because these settings call for nonsmooth regularizations. We tackle this problem by including constraints on the spatial variations and value ranges of the inverted velocities, as opposed to adding penalties to the objective, which is more customary in mainstream geophysical inversion. By demonstrating the lack of predictability of edge-preserving inversion when the regularization is in the form of an added penalty term, we advocate the inclusion of constraints instead. Our examples show that the latter leads to more predictable results and to significant improvements in the delineation of salt bodies when these constraints are relaxed gradually in combination with extending the search space to approximately fit the observed data but not the noise.