Oil and gas is a huge industry in the United States, and is currently experiencing a boom in the Permian Basin. This oil-rich region stretches from western Texas to eastern New Mexico. Each day, hundreds of new well pads appear across the landscape, making it difficult for regulators to keep up with. But unregistered well pads are both a safety hazard and a missed opportunity for revenue for agencies such as the Bureau of Land Management. To help regulators monitor the progress of new drilling on their land as well as look for potential illegal drilling, you’ll use the deep learning capabilities of the ArcGIS API for Python and ArcGIS Pro. The full workflow, from exporting training data and training a deep learning model to detecting objects across a large landscape, can be done in the Python API. Rohit’s Jupyter notebook from the Developer Summit Plenary session shows how he did this with the help of the arcgis.learn module.
In the notebook, Rohit uses a set of training data to train a deep learning model to recognize well pads. He then deploys the model with the computing power of his portal to identify well pads in the New Mexican desert.
To train the model, Rohit uses the SingleShotDetector model, so called because it’s able to find all objects within an image in one glance. This model can be saved as an Esri Model Definition that can be used with ArcGIS Pro, or a Deep Learning package that can be used for inferencing using Raster Analytics. The Detect Objects Using Deep Learning tool is now available with Image Server and allows distributed inferencing at scale using GPUs.
For further analysis and visualization, you can pull the well pad layer into ArcGIS Pro. In his portion of the demo, Vinay compares a layer showing all the permits that have been issued in the region to Rohit’s well pad results.
These results, once verified, can be used to direct ground crews or inspectors to the most critical sites. You can extend this workflow using Workforce for ArcGIS, Operations Dashboard, Web App templates, and more.