Image processing for the oil sands mining industry [In the Spotlight]

摘要

Oil sands mining is an outdoor and continuous operation, conducted under all weather conditions. Robust and reliable image processing algorithms are called for to deliver accurate information with which operational decisions are made. Oil sands are observed at many points in the ore preparation pipeline, and each scenario defines a separate problem and presents different challenges. For example, oil sands can be imaged at the entrance to the crusher, on a conveyor belt after crushing, or before or after screening on a largely empty belt or with the large fragments amid fine fragments. The problem of measuring ore size can be formulated as one of image segmentation in which the foreground - i.e., relatively large ore fragments - are to be identified and delineated from each other as well as from the background, in a static image or in a video sequence. In what follows we describe the scenario and main steps of the image segmentation-based solution.