Improved geological mapping using Landsat TM data, Southern Red Sea Hills, Sudan: PC and IHS decorrelation stretching

Abstract

As part of an on-going research in geological applications of remote sensing over the Red Sea Hills, Sudan, different image processing algorithms have been evaluated in context of geological mapping using Landsat TM data. IHS and Principal Component (PC) decorrelation stretching, among others, appear to provide the maximum geological information. The major basement rocks in the area include gneisses, varying compositions of metavolcano-sediments, mafic to acidic synorogenic plutons, and dykes. Basalt, rhyolite, and sandstone make the major Phanerozoic cover. IHS-transformed images, with substantially saturated colour, appear generally superior in discriminating the various rock units when three least correlated bands are used. Images obtained from PC decorrelation stretching, computed with reduced noise, have been more informative in distinguishing lithologies with subtle compositional difference. Structural information are well exhibited on both images and undocumented shear zones, prominent faults and lineaments of regional significance have been detected thereby improving the geological understanding of the area. The study testifies that using Landsat TM data PC and IHS decorrelation-stretching methods yield the best results for geological mapping in arid regions, by preserving morphologic and spectral information, and when combined, they can be very helpful for improvements in already mapped areas.