Comparison of Support Vector Machine and Artificial Neural Network for Delineating Debris Covered Glacier

Glacier mapping accuracy plays very important role in studies like mass balance of glacier, water resource management and in understanding the health of the glacier. Several of the present glaciers are covered with debris of different thickness. So it becomes difficult to distinguish debris covered glacier from the adjacent valley rock, alone with the use of optical data because of the same reflectance in visible to near infrared region. In this paper we have trained Support vector machine (SVM) and Artificial neural network (ANN) on several parameters such as slope, surface curvature, thermal data and also on several texture parameter, such as variance, skewness, entropy, homogeneity, mean and dissimilarity. Then both the algorithms were applied on the part of the alaknanda basin. It was observed that both ANN and SVM produced good results, with accuracy of SVM slightly higher than that of ANN algorithm.