Labelled data for P/S separation with CNN
收藏资源简介：
This is the part of training labelled Dataset-B for the publication "P/S separation of multi-component seismic data at land surface based on deep learning". There are 500-shot data labels: Train/, Val/ & Test/ are the separated file for Training, Validation & Testing The labelled data size are nx*nz=1001*3001 using a IEEE float format One can use the Seisic Unix command to plot for QC: ximage < shotx_mod_1.dat n1=3001 perc=99 & In each directory, the files are named as follows: The horizontal component: shotx_mod_*.dat The vertical component: shotz_mod_*.dat The P-wave label : shotp_*.dat The S-wave label : shots_*.dat
本数据集为发表论文《基于深度学习的陆地地表多分量地震数据P/S波分离》中标记训练数据集-B的一部分。本数据集包含500炮带标注数据，其中Train/、Val/与Test/分别为训练、验证与测试所用的拆分文件。标注数据的尺寸为nx*nz=1001*3001，采用IEEE浮点格式存储。可通过Seisic Unix命令绘图以完成质量控制（QC）：ximage < shotx_mod_1.dat n1=3001 perc=99 &。各子目录内的文件命名规则如下：水平分量文件为shotx_mod_*.dat；垂直分量文件为shotz_mod_*.dat；P波标签文件为shotp_*.dat；S波标签文件为shots_*.dat。
