Emerging Trends in the Mapping of Contaminated Snow Using Hyperspectral Images and Field Spectra

Abstract

In this study, glaciers of the Patsio of Himachal Pradesh, India, are selected. The field spectra were collected for pure snow, contaminated snow, and adjacency factors using a Spectro-radiometer. Contaminants, like ash, coal, salt, sand, soil, wood, and metals are mixed with snow in the field in varying amounts (2.5g and 5g). Similarly, the spectra of green vegetation, vegetation covered with snow, wet grass, etc., were acquired. The authors also demonstrated the utility of AVRIS-NG hyperspectral data in linear spectral analysis to classify spectrally similar objects. The preliminary results from the linear unmixing technique for pure snow and its respective mixed components, like snow with debris, and frozen lakes are synchronized with the Google Earth imagery and are acclaimed for their validity. Contaminated snow analysis was done using the normalized difference snow index and S3-based snow indexes. Snow under vegetation and areas with snow cover could be mapped using both snow indexes.