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/

Global: Zendy Plus annual

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

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

In the audio magnetotellurics (AMT) sounding data processing, the absence ofsferic signals in some time ranges typically results in a lack of energy in theAMT dead band, which may cause unreliable resistivity estimate. We propose adeep convolutional neural network (CNN) to automatically recognize sfericsignals from redundantly recorded data in a long time range and use them tocompensate for the resistivity estimation. We train the CNN by using field timeseries data with different signal to noise rations that were acquired fromdifferent regions in mainland China. To solve the potential overfitting problemdue to the limited number of sferic labels, we propose a training strategy thatrandomly generates training samples (with random data augmentations) whileoptimizing the CNN model parameters. We stop the training process and datageneration until the training loss converges. In addition, we use a weightedbinary cross-entropy loss function to solve the sample imbalance problem tobetter optimize the network, use multiple reasonable metrics to evaluatenetwork performance, and carry out ablation experiments to optimally choose themodel hyperparameters. Extensive field data applications show that our trainedCNN can robustly recognize sferic signals from noisy time series for subsequentimpedance estimation. The subsequent processing results show that our methodcan significantly improve S/N and effectively solve the problem of lack ofenergy in dead band. Compared to the traditional processing method withoutsferic compensation, our method can generate a smoother and more reasonableapparent resistivity-phase curves and depolarized phase tensor, correct theestimation error of sudden drop of high-frequency apparent resistivity andabnormal behavior of phase reversal, and finally better restore the realshallow subsurface resistivity structure.

Deep learning based sferics recognition for AMT data processing in the dead band