Full text loading...
We present AI-drive workflow for automatic interpretation of large seismic volumes. The workflow is applied to an offshore Abu Dhabi area of 15000 km2. First, we show the fault detection results for the low-magnitude faults of strike-slip type by the fine-tuning of the DNN. The fault detection is conducted in a comprehensive way with structure-enhancing denoise for feature preservation. Second, we illustrate that automated horizon interpretation and flattening allows us to observe a variety of geological features, which are difficult to target otherwise. Finally, we show how the structure-enhancing denoise improves the geological feature identification by an example of channel detection. We further plan to apply the suggested workflow for a larger data set covering the whole Abu Dhabi offshore region.