Development of an automatic technique using satellite data and DEM for identification of landforms in different geomorphic provinces of India
Understanding the landforms present in different geological provinces like Fluvial, Glacial, Coastal, Denudational, Aeolian etc based on terrain analysis and classification of precise DEM using various satellite and aerial data sets. Exploring the semi-automatic and automatic techniques like Machine based methods, Artificial Neural Network, Object Based classification, Fuzzy logics etc. for automatic classification of the landforms by reducing the manual delineation efforts and time as well as to support the modeling of processes as hydrological modeling, geomorphic modeling etc.
Numerical modelling for crustal deformation using GPS vectors
NRSC operated Continuously Operated Reference stations in 6 locations within the HP and Uttarakhand Himalayas. Differential movement as recorded by GPS observations indicate towards active convergence in the region. Further, as a resultant of such differential movement, it may be possible, that there is significant strain accumulation in the region, pervasively or localized to certain areas. These areas are then prone to future earthquakes. To understand the state of strain in the area, estimation and modeling of the same is important. For this, numerical models (Finite Element) are to be created to ingest the GPS vector data, the fault geometries with the crustal rheological properties and simulate the strain accumulation in the region
Derivation of innovative sub-pixel mapping algorithms to identify rarely occurring class indicating mineralisation
Under this research domain, focus should be on the derivation of the effective subpixel mapping method which may help to identify the spectrally diagnostic target which occupies 20 to 30 percent of the pixel but. Algorithms should be effective for processing both hyperspectral and multispectral data. These algorithms should also account to characterize complex (stochastic) character of background
