Optimizing mineral exploration in resource constrained environments through satellite-based structural geology mapping

This study assessed the performance of Landsat 8, Sentinel 2 A, PlanetScope, Gaofen-1 and SRTM-DEM in extracting geological lineaments for gold mineral exploration. Principal Component Analysis (PCA), Minimum Noise Fraction (MNF), Independent Component Analysis (ICA), and Band Rationing (BR) were applied on satellite images to automatically extract lineaments. Then, an overlay analysis was performed to compare the extracted lineaments with those from geological bulletins. MNF outperformed PCA, ICA and BR for Landsat and Sentinel images in terms of lineament direction with most lineaments oriented in S-N and SE-NW directions. Spatial resolution was the major determinant of lineament number and length which in turn influenced lineament density. The high spatial resolution imagery, i.e. Gaofen-1 and PlanetScope had the highest lineament density and number of extracted lineaments (i.e. 50000 and 80000) and the least on the coarser ones (Landsat and SRTM-DEM with lineaments <25000). The percentage of spatial overlap between geological faults derived from geological bulletins and satellites was less than 8%, 3% and 2% for satellites with a 30 m,10m and 5 m spatial resolution, respectively. There was considerable overlap between gold mineralization points and lineaments, especially in high lineament density areas suggesting the utility of the developed approach for mineral exploration.