junlysky/PhaseNet

Related paper:

Weiqiang Zhu, Gregory C Beroza; PhaseNet: a deep-neural-network-based seismic arrival-time picking method, Geophysical Journal International, Volume 216, Issue 1, 1 January 2019, Pages 261–273, https://doi.org/10.1093/gji/ggy423

1. Install

The code is tested under Python3.6.

Using virtualenv

pip install virtualenv
virtualenv .venv
source .venv/bin/activate
pip install -r requirements.txt

Using Anaconda

conda create --name venv python=3.6
conda activate venv
conda install tensorflow=1.10 matplotlib scipy pandas tqdm

2.Demo Data

Located in directory: dataset

3.Model

Located in directory: model/190703-214543

4. Prediction

Data format

Required a csv file and a directory of npz files.

The csv file contains one column: "fname"

The npz file contains one variable: "data"

The shape of "data" variable has a shape of 3000 x 3

source .venv/bin/activate
python run.py --mode=pred --model_dir=model/190703-214543 --data_dir=dataset/waveform_pred --data_list=dataset/waveform.csv --output_dir=output --plot_figure --save_result --batch_size=20

Notes:

1. For large dataset and GPUs, larger batch size can accelerate the prediction.
2. Plotting figures is slow. Removing the argument of --plot_figure can speed the prediction
3. If using input data length other than 3000, specify argument --input_length=. But this is not suggested as the model is trained using input length of 3000. Too long input length would degrade the performance.
4. The activation thresholds for P&S waves are set to 0.5 as default. These two values can be changed to improve the detection performance. Specify --tp_prob and --ts_prob to change the two thresholds.

5. Train

Data format

Required a csv file and a directory of npz files.

The csv file contains four columns: "fname", "itp", "its", "channels"

The npz file contains four variable: "data", "itp", "its", "channels"

The shape of "data" variables has a shape of 9001 x 3

The variables "itp" and "its" are the data points of first P&S arrivals picked by analysts.

source .venv/bin/activate
python run.py --mode=train --train_dir=dataset/waveform_train --train_list=dataset/waveform.csv --batch_size=20

6. Valid and Test

source .venv/bin/activate
python run.py --mode=valid --model_dir=model/190703-214543 --data_dir=dataset/waveform_train --data_list=dataset/waveform.csv --plot_figure --save_result --batch_size=20

Please let us know of any bugs found in the code. Suggestions and collaborations are welcomed!