Neural network analysis of comprehensive exploration data to assist in brown-field exploration for an Australian mid-tier mining company

A multi-variate unsupervised and supervised neural network study was successfully undertaken on comprehensive advanced exploration data confirming the location and extent of known mineral occurrences as well as identifying additional exploration targets. The dataset comprised space-borne, airborne and ground-based sampling techniques including Landsat 8, ASTER, SRTM, aeromagnetics, radiometrics, transient electromagnetics, gravity, multi-element bedrock geochemistry and depth to bedrock. Study provided products which significantly improved the lithostructural mapping of the bedrock (largely under cover) as well as identified additional features having similar component profiles as known deposits and prospects. The study involved the use of a variety of software applications including Global Mapper, Surfer, QGIS and a proprietary neural network system.