Glacier Boundary Mapping Using Deep Learning Classification over Bara Shigri Glacier in Western Himalayas

Glacier, snow, and ice are the essential components of Himalayan cryosphere provide a sustainable water source for different applications. Continuous accurate monitoring glaciers allows forecasting analysis natural hazards resource management. In past literature, methodologies such as spectral unmixing, object-based detection, combination various indices commonly utilized mapping ice, glaciers. Most these methods require human intervention in feature extraction, training models, validation procedures, which may create bias implementation approaches. this study, deep learning classifier based on ENVINet5 (U-Net) architecture is demonstrated delineation glacier boundaries along with snow/ice over Bara Shigri (Western Himalayas), Himachal Pradesh, India. Glacier Landsat data takes advantage long coverage period finer spectral/spatial resolution wide larger scale. Moreover, utilizes semantic segmentation network to extract boundaries. Experimental outcomes confirm effectiveness (overall accuracy, 91.89% Cohen’s kappa coefficient, 0.8778) compared existing artificial neural (ANN) model 88.38% 0.8241) generating classified maps. This study vital cryosphere, hydrology, agriculture, climatology, land-use/land-cover analysis.