Dataset formation for sem-based images of commercial minerals using ml algorithms: case study for zircon in the Egyptian mountains

Abstract Mineral identification holds paramount importance in geological and mineralogical endeavors, encompassing exploration, mining, and mineral processing. This work underscores the time-consuming and equipment-dependent nature of conventional identification methods, advocating for the integration of artificial intelligence techniques, particularly machine learning and computer vision. Commercial minerals, including zircon, are identified as linchpins of various industries, particularly ceramics and dentistry. The work elaborates on the pivotal role of SEM imaging techniques in discerning economic minerals in granitic rocks and pegmatite, emphasizing their utility in environmental science and mineral exploration. A novel computational approach is introduced, offering automation of mineral grain recognition, thereby mitigating the laborious and resource-intensive aspect of the process. The subsequent discussion pertains to the creation of a specialized SEM image dataset focusing on Egyptian commercial minerals, commencing with zircon, a dataset with foreseeable extensions. The authors anticipate that this dataset will significantly contribute to mineralogical research, facilitating precise mineral identification through AI techniques and enriching insights into Egypt’s geological wealth.