Deep learning and fractal-wavelet techniques for magnetite-apatite exploration in Tarom Iran

This study proposes an approach to detect iron-apatite mineralization and paragenesis related to Iron deposits by applying an image fusion based on deep learning (FDL) on exploration datasets. The exploration dataset of the study area includes remote-sensing, airborne geophysics, stream sediment geochemistry, and the 1:100,000 geological map of the Tarom area (Iran). The dataset used as a layer or sublayer in the Geographic Information System (GIS) program and exploration model includes features based on the presented Prospectivity Model (PM) of Iron-Apatite. The Prospectivity model was based on logical relations between each layer or sublayer extracted features and classes. The thresholds were performed on the layers and sublayers based on fractal-wavelet algorithms (DWT, W-N, SA) using the fractal wavelet classifier tool (FWC). To indicate the potential regions as Mineral Prospectivity Map (MPM), the results of each layer are integrated based on exploration model relations in a raster format by pixel values in application of fusion-based deep Learning (FDL) algorithm using pre-trained networks.