Copyright (c) 2022 Zhengjie Zhang (zhangzhengjie@mail.ustc.edu.cn)
- This is the first part of the subject Analysis and association of vehicle signals in multi-station seismic data based on artificial intelligence.
- The second part of the subject belongs to the content of using convolutional neural networks to associate seismic signals. At present, the code is still being sorted out and has not been made public.
- The published data belongs to the test data.
- K-Shape: K-Shape algorithm is used for clustering analysis to find the similarity of signals.
- xcorr: The cross-correlation test of the synthetic data and the cross-correlation results of the received signals of the actual stations.
- ObsPy-Tutorial.pdf: ObsPy Chinese tutorial -V 1.0 (2020/04/12), you can get more detailed use of ObsPy through the official website https://docs.obspy.org/
-
Create a new python virtual environment
MLMSA
and activate itconda create -n MLMSA python=3.8 conda activate MLMSA
-
Install the package
(Note: We do not recommend that you install the package below the following version, the higher version is adapted.)
conda install numpy==1.21.5 pandas==1.4.2 matplotlib==3.5.1 scipy==1.8.0 obspy==1.3.0
If your installation fails, you can try to replace
conda
withpip
. In addition, you can also try again after replacing Tsinghua or Ustc sourceconda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/ conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge/ conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/msys2/ conda config --set show_channel_urls yes
or
conda config --add channels https://mirrors.ustc.edu.cn/anaconda/pkgs/main/ conda config --add channels https://mirrors.ustc.edu.cn/anaconda/pkgs/free/ conda config --add channels https://mirrors.ustc.edu.cn/anaconda/cloud/conda-forge/ conda config --add channels https://mirrors.ustc.edu.cn/anaconda/cloud/msys2/ conda config --add channels https://mirrors.ustc.edu.cn/anaconda/cloud/bioconda/ conda config --add channels https://mirrors.ustc.edu.cn/anaconda/cloud/menpo/ conda config --set show_channel_urls yes
- Set your working directory at
/data/
cd /data/ git clone https://github.com/zhangzj1209/MLMSA.git unzip MLMSA.zip cd MLMSA/
cd K-Shape/
python K_Shape.py
If you want to modify some basic parameters, it will be in ./config_KShape.py
, and the parameters of the clustering algorithm are in the main function part of ./K_Shape.py
.
The logging of K-Shape results are recorded in ./log.txt
, loss function is recorded in ./Loss.npy
, and the number of clusters is recorded in ./NUM_CLU.npy
.
Then, you can use ./plot_KShape
to show related figures.
cd xcorr/
- run the example of synthetic data
python synthetic_data.py
- run the result of field data
If you want to modify some basic parameters, it will be in
python xcorr.py
./config_xcorr.py
.