Combining a digtial elevation model and thermal information for automated glacier mapping

We combined the morphometric parameters such as elevation, slope information and the thermal band to map glacier by analyzing the location of glaciers and debris-covered glaciers and the difference temperature between supraglacial debris and the adjacent periglacial debris outside the glacial margins. Support Vector Machine (SVM) was used to classify the combined image. Confusion matrices and the glacier outline derived from visual interpretation were used to evaluate the effectiveness of the employed method in this study. The overall accuracy with 91.75% and 94.05% of the TM3/TM5 ratio and the employed method respectively indicated the combination of topography features and the thermal band can delineated the boundary of the glacier more precisely and this method is promising for glacier mapping.