Statistical Models for Retrieval of Himalayan Land Surface Parameters for Geomorphologic ResearchJournal: International Journal of Science and Research (IJSR) (Vol.2, No. 3)
Publication Date: 2013-03-05
Authors : Manjeet Singh; V. D Mishra; N. K Thakur; Jyoti Dhar Sharma;
Page : 330-340
Keywords : AWiFS; Hyperion; FLAASH; SAM; MLH; SVM;
Qualitative and quantitative estimation of land surface parameters has great concern for ecological and hydrological system. This makes land surface parameters as an important tool to study earth�s climate system especially when satellite data provide timely and efficient information about large land area. In the present paper, the study was carried out by using NASA�s hyper spectral EO-1 Hyperion sensor and multi-spectral Advance Wide Field Sensor (AWiFS) of IRS-P6 for different ranges (lower, middle and upper) of Himalaya. The analysis procedure consists of Fast Line-of-sight Atmospheric Analysis of Spectral Hyper cubes (FLAASH) atmospheric correction code derives its physics-based algorithm from the Moderate Resolution Transmittance (MODTRAN4) radiative transfer code as well as radiometric (atmospheric + topographic) correction to retrieve surface reflectance. The terrain characteristics have been extracted from Digital Elevation Model (DEM) using 1:50, 000 scale SoI maps at 40m contour interval. Various statistical models for supervised classification such as spectral angle mapper (SAM), support vector machine (SVM), and maximum likelihood (MLH) has been examined and validated with existed Normalized Difference; Vegetation Index (NDVI), Snow Index (NDSI) and Glacier Index (NDGI) models. The spectral reflectance of different surface parameters has been collected in field, using spectro-radiometer and compared with satellite derived spectra. Present work has focused on three key issues (a) accurate registration of the images for land cover maps (b) estimation of spatial distribution of snow cover at sub pixel level and (c) multi-temporal input to hydrological, ecological and land surface modeling. Study distills these statistical approaches into a unique set of hierarchical taxonomy that reveals the similarities and differences between algorithms.
Other Latest Articles
Last modified: 2021-06-30 20:14:29