Automated lineament mapping from remotely sensed data: case study Osun drainage basin, southwestern Nigeria
The automatic procedures of PCI LINE and Imagine Objective Line Extraction were adopted to extract lineaments from Landsat OLI imagery (band 6) and Digital Elevation Models (SPOT DEM and ASTER DEM) of Osun Drainage Basin, Southwestern Nigeria. This was with a view to optimally map lineaments within the basin. The optical images were initially subjected to Median filtering in order to edge out all unwanted linear information while preserving the target geotectonic lineaments. Thereafter, the filtered images were subjected to both linear (directional) and non-linear (non-directional) edge enhancement algorithms. The resultant omnidirectional and non-directional edge enhanced images were subjected to Object-based Image Analyses, which culminated to lineament map of high accuracy. Subsequently, the resultant composite lineaments were categorized into morphological and geotectonic lineaments based on lithological and hydrological information. Furthermore, borehole data were employed to assess the hydrogeological significance of the geotectonic lineaments. It was observed that the number of lineaments increased with decreased azimuth interval of filtering directions with SPOT DEM giving the highest number of lineament. Directional filtering along N-S and E-W directions proved to be the most efficient in detecting most lineaments within the study area. Forty seven per cent of the composite lineaments were adjudged to be hydrogeologically significant with dominant azimuth directions of NE-SW and NW-SE while minor trends (N-S and E-W) were also represented. The clustering of hydrolineaments and the corresponding lineament intersections coincided with the occurrence of felsic rocks particularly within the upland axis of the study area. Results of this study suggested that the more the input data and the adopted techniques, the higher and more reliable are the resultant lineaments. Moreover, SPOT DEM proved to be the most efficient among the input optical datasets.
Keywords: Remote Sensing, Lineaments, Hydrogeology, Image Processing, Edge Enhancement