Seismic surveys for windfarm projects are shot at ultra-high-resolution to better image shallow sediments that are critical for installing large wind turbines. The size and complexity of these datasets can make them difficult to interpret and analyze quickly. Pre-trained machine learning models struggle to produce optimal predictions since they were built using seismic responses at the E&P scale. However, we can apply an interactive deep learning methodology to guide the neural network to capture shallow geological features on these ultra-high-resolution datasets to determine optimal position for at-scale offshore wind turbine installation.
Bluverse Case Study: Deep Learning on Shallow Facies & Hazard Characterization for Offshore Wind
