Principal components analysis for borate mapping

Abstract

Principal components analysis (PCA) of remotely sensed satellite image data is a widely used method in mineral exploration. Generally, the method is used for iron oxide and hydroxyl mapping. In this study, however, the PCA method is adopted for borate exploration. This paper demonstrates how PCA of Landsat TM data can be used to map borate minerals. The method has been applied to the sub-scene of Bigadic and tested on the borate field in Kirka, Turkey. Anomalous pixels for borate minerals in PC6 images have coincided with known borate deposits. Whether borate minerals are mapped into a PC image depends on the appearance of opposite signs in eigenvector loadings for TM4 and TM7 in one or more PCs. Borate coverage in an image is important to emphasize the appearance of opposite signs in eigenvector loadings for TM4 and TM7 in more than one PC.