The nature of statistical learning theory
The nature of statistical learning theory
International Journal of Remote Sensing
A comparison of methods for multiclass support vector machines
IEEE Transactions on Neural Networks
A self-trained semisupervised SVM approach to the remote sensing land cover classification
Computers & Geosciences
| Hi-index | 0.00 |
Remote sensing data can be effectively used as a means to build geological knowledge for poorly mapped terrains. In this study, the support vector machine (SVM) algorithm is applied to an automated lithological classification of a study area in northwestern India using Advanced Space-borne Thermal Emission and Reflection Radiometer (ASTER) imagery, together with ASTER-derived digital elevation model (DEM) and aeromagnetic data. Image enhancement techniques were used to produce derivative datasets from those three datasets to improve lithological discrimination. A series of SVMs were tested using various combinations of input datasets selected from among 47 datasets including the original 14 ASTER bands and 33 derivative datasets extracted from the ASTER, DEM and aeromagnetic data, in order to determine the optimal inputs that provide the highest classification accuracy. A combination of ASTER-derived independent components, principal components, DEM-derived slope, curvature and roughness, and aeromagnetic-derived mean and variance of magnetic susceptibility provided the highest overall classification accuracy of 92.34% for lithological classes on independent validation samples. Comparison with maximum likelihood classifier (MLC) show that the SVM provides higher accuracy both in terms of classification of independent validation samples as well as similarity with the available bed-rock lithological map. The study illustrates that SVM can help in building first-pass lithological map for areas for which some information on the types of lithologies present is available.