English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Zendy Tools

Zendy Open

Seismic velocity is one of the most important parameters used in seismicexploration. Accurate velocity models are key prerequisites for reverse-timemigration and other high-resolution seismic imaging techniques. Such velocityinformation has traditionally been derived by tomography or full-waveforminversion (FWI), which are time consuming and computationally expensive, andthey rely heavily on human interaction and quality control. We investigate anovel method based on the supervised deep fully convolutional neural network(FCN) for velocity-model building (VMB) directly from raw seismograms. Unlikethe conventional inversion method based on physical models, the superviseddeep-learning methods are based on big-data training rather thanprior-knowledge assumptions. During the training stage, the network establishesa nonlinear projection from the multi-shot seismic data to the correspondingvelocity models. During the prediction stage, the trained network can be usedto estimate the velocity models from the new input seismic data. One keycharacteristic of the deep-learning method is that it can automatically extractmulti-layer useful features without the need for human-curated activities andinitial velocity setup. The data-driven method usually requires more timeduring the training stage, and actual predictions take less time, with onlyseconds needed. Therefore, the computational time of geophysical inversions,including real-time inversions, can be dramatically reduced once a goodgeneralized network is built. By using numerical experiments on syntheticmodels, the promising performances of our proposed method are shown incomparison with conventional FWI even when the input data are in more realisticscenarios. Discussions on the deep-learning methods, training dataset, lack oflow frequencies, and advantages and disadvantages of the new method are alsoprovided.

Deep-learning inversion: a next generation seismic velocity-model building method