Fault Detection using Convolutional Neural Network

Fault delineation is one key of the reservoir characterization. By interpret fault delineation, we can know is there any potential hydrocarbon trap and hydrocarbon migration. By doing this manually sometimes it can lead into miss-interpretation and not efficient. So in this case, I used Deep Learning model which is Convolutional Neural Network with UNet architecture to predict where is the fault delineation in my seismic data.

Also I do the comparison with established and powerful fault delineation method which is Ant-track. The result shows, the CNN model produce more clearly the fault delineation and also faster time computation than the Ant-track method.

Skills: Python, Deep Learning, Segyio, Numpy, Matplotlib, TensorFlow, Keras, Scikit-Learn, Mayavi