Fig. 6: Overall tiling, well pad detection, verification, and storage tank detection process.
From: Deep learning for detecting and characterizing oil and gas well pads in satellite imagery
a The deployment region was tiled into smaller images, b each of which were input to the well pad detection model which produced bounding boxes indicating the location of potential well pads. c Images around the candidate well pads were verified if they matched reported data or if they were confirmed by the verification model. d Finally, images around the verified well pads were input to the storage tank detection model to identify the locations of any storage tanks on the pads.
