In this paper, we present an application of most advanced DL (Deep Learning) approaches, in characterizing subsurface salt body on seismic images. Nowadays, DL has been widely used in many industries and achieving tons of successful results, while it is pioneering work to explore efficient application of DL in automation of geological interpretation. In our case study research, we utilized end to end semantic segmentation technology to automatically identify salt body, based on the data provided by TGS in the Kaggle Platform. The key to our success is the combination of state-of-art architecture and model ensembling. Based on the high IoU (Intersection over Union) score [Rahman] [Nowozin] from our final model in evaluating the test data, we demonstrate the strength of DL in Salt Body auto-detection, and further identify DL as a trending technology in oil/gas exploration.

Presentation Date: Wednesday, September 18, 2019

Session Start Time: 1:50 PM

Presentation Time: 3:05 PM

Location: Poster Station 2

Presentation Type: Poster

This content is only available via PDF.
You can access this article if you purchase or spend a download.