Abstract
Efficiently addressing challenges in mapping spectrally similar classes and overcoming limitations imposed by the restricted spectral sensitivity of remote sensing data is crucial for accurate image processing. In this context, the significance of the Maximum Likelihood Classifier (MLC) becomes apparent in optimizing classification accuracy. This study focuses on classifying glacierized terrain surrounding the Kolahoi glacier in the Kashmir Valley, integrating a diverse set of data sources, including Digital Elevation Model (DEM)-derived topographic attributes and transformed spectral data. The integration of ancillary sources, such as slope, aspect, plane, and profile curvatures derived from DEM, along with spectral information extracted from the LISS-III image using Red/SWIR ratio, NIR/SWIR ratio, and Normalized Difference Snow Index (NDSI), is explored to classify nine distinct glacier terrain classes. These classes encompass a range of features, including snow, ice, mixed-ice, debris, supraglacial debris, periglacial debris, water, vegetation, bare land, and shadow.To optimize the classification process, various scenarios are examined by combining each ancillary data source with the four spectral bands. The MLC is employed as the classification algorithm, and the resulting glacier terrain maps are validated against an ASTER image of Area of Interest (AOI). Notably, the study reveals that the glacier terrain map achieving the highest overall accuracy, an impressive 90.75%, emerges when incorporating the Red/SWIR ratio with the spectral data. This outcome underscores the effectiveness of the proposed classification approach in accurately mapping diverse glacier terrain classes.The successful application of the Maximum Likelihood Classifier in this study holds promising implications for applications such as glacier monitoring and environmental assessment in glacierized regions. The optimized classification approach presented here demonstrates its potential to contribute significantly to remote sensing applications, offering a valuable tool for understanding and monitoring complex terrain features in glacierized environments.