Dr. Saurabh Khanna, Electronics and Communication Engineering
IIT Roorkee
coordinator_spacetech@iitr.ac.in
CO-Principal Investigator
Shri Indranil Mishra, SIPA-ODPG-ODPD, SAC
IIT Roorkee
Project Overview
The main objective of the proposed work are
1. Posing the denoising and super-resolved unmixing of hyperspectral images (HSIs) from hyperspectral measurements as desketching of an r-separable matrix from its compressive and subsampled measurements.
2. Python-based proof-of-concept demo of denoising and super-resolved hyperspectral image reconstruction (including unmixing) from point spread function (PSF) hyperspectral measurements. The joint denoising and unmixing algorithms will rely on r-separable modeling of the HSIs.
3. GPU-based acceleration of computationally intensive steps of the proposed joint denoising and unmixing algorithms for faster runtimes. PyTorch library will be used to implement the underlying tensor computations on GPU hardware.
4. Designing optimized strategy to a priori select the frequency bands for hyperspectral measurements for a given application-specific material spectra library.
Achievements
The project being undertaken by ISRO in collaboration with academia is directly or indirectly beneficial to the society.