Seismic Fault Interpretation Using 3-D Scattering Wavelet Transform CNN
Shian Shen, Haishan Li, Wenchao Chen, Xiaokai Wang, Binke Huang
Abstract
Fault interpretation is very important for reservoir characterization in seismic petroleum exploration. Recently, different machine learning methods have been widely performed in fault detection of seismic data. Faults have multi-scale characteristics and present abrupt changes in seismic data. Based on this characteristic, we introduce a three-dimensional scattering wavelet transform (3D-SCWT) CNN (convolutional neural network) method, called 3D-SCWTnet, which contains the 3D-SCWT and the 3D-Unet. 3D-SCWT has multi-scale and multi-direction characteristics to delineate the spatial characteristics at different scales and angles of the faults in seismic data. 3D-Unet is the three-dimensional CNN for the faults classification task. The 3D-SCWTnet makes full use of the multi-scale properties of 3D-SCWT and deep learning network. The synthetic seismic data is used as the training data for the 3D-SCWTnet, the predicted results of synthetic data and field seismic cube obtain a higher accuracy than the traditional 3D-Unet.