Well Pad Detection

Oil and gas sector development in the USA has shown significant growth over the last decade. The reasons for that includes technological progress in the industry, advanced techniques used for hydrocarbon extraction, and new sources of hydrocarbon recovery that were previously unreachable for the industry. Innovations such as horizontal drilling, hydraulic fracturing, advanced seismic imaging, and modern field development techniques allow increasing the production capabilities on numerous oil- and gas-bearing sites in the country. In particular, the Permian basin is developing very rapidly, which results in continuous expansion of drilling operations and related infrastructure.

When the production of oil and gas grows, the number of construction works of well pads increases as well. A well pad is an area where all necessary work is done for exploration, drilling, and extraction of hydrocarbons. It involves clearing the area, leveling it for use as infrastructure, and constructing additional infrastructure required for drilling and extraction operations. Monitoring and detection of well pads is necessary for the identification of changes in energy development and monitoring of new drilling activities.

This deep learning model will be used for automatic detection and classification of oil and gas well pads on the images taken by satellites or by aerial photography. Using artificial intelligence and computer vision technology will enable fast and efficient detection and monitoring of well pads across large areas.

Such a model has various applications, including energy market intelligence, competitor analysis, exploratory monitoring, regulatory compliance, land use evaluation, and infrastructure planning. Companies can leverage well pad locations identified by the model to monitor new drillings, evaluate shifts in production areas, and get a better picture of expansion through time.

Additionally, automatic identification of well pads can be used for regulatory monitoring and can assist analysts in detecting the presence of any illegal or unauthorized drilling. Once combined with historical imagery, geographic information system data, production numbers, and other geospatial data, the findings of such research will enable companies to gain valuable information about oil and gas infrastructure development.

Thus, by transforming large amounts of imagery into actionable geospatial data, the presented deep learning model enables efficient and effective monitoring of well pads in energy production areas.