Spectral-Spatial

This repository contains code for hyperspectral image denoising using TV based regularizers. Optimisation is performed using Chambolle-Pock algorithm [1]

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/spectral-spatial-analysis.git
   cd spectral-spatial-analysis

2. Install the required dependencies:

   pip install -r requirements.txt

Usage

Algorithms

• chambolle_pock.py: Implementation of the Chambolle-Pock algorithm.
• grad_alignement.py: Implementation of denoising using a gradient alignment regularization.
• nabla.py: Contains functions for computing gradients.
• tvprior.py: Implementation of denoising using a Total Variation Prior.
• tv_plus_grad_alignement.py: Combines Total Variation and gradient alignment.

Datasets

• datasets.py: Functions for loading and processing datasets.
• show_dataset.ipynb: Jupyter notebook for visualizing datasets.

Metrics

• metrics.py: Functions for computing various metrics.

Scripts

• ACP.ipynb: Jupyter notebook for ACP analysis.
• benchmark.py: Script for benchmarking algorithms.
• grad_alignement.ipynb: Jupyter notebook for denoising using a gradient alignment regularization.
• HyDe.ipynb: Jupyter notebook for HyDe analysis.
• tv_plus_grad_alignement.ipynb: Jupyter notebook for TV + Gradient Alignment.
• tvprior.ipynb: Jupyter notebook for denoising using a TV Prior.

Results

• Contains results from various experiments and analyses.