Nonlocal Low-Rank Tensor (NLRT)

This repository contains the MATLAB code for the paper Hyperspectral Image Reconstruction of SD-CASSI Based on Nonlocal Low-Rank Tensor Prior.

All test data are available at this link on Google Drive.

Figure 1. Reconstruction process of the hyperspectral image for scenario S3 as described in the corresponding paper.

Figure 2. Reconstruction process of the hyperspectral image for scenario S4 as described in the corresponding paper.

Usage

Download the NLRT repository

0. Requirements are MATLAB(R) with Parallel Computing Toolbox (parfor for multi-CPU acceleration).
1. Download this repository via git

git clone https://github.com/sdnjyxr/NLRT

Run NLRT on hyperspectral images

3. Test the NLRT algorithm via

NLRT.m

Structure of directories

directory	description
datasets	data used for reconstruction
funs	functions of main algorithms
results	reconstruction results
tools	the toolbox used by the main algorithm
utils	utility functions

Platform

The testing environment utilizes MATLAB(R) R2023a on Ubuntu 22.10, equipped with an Intel® Xeon® w7-2495X × 48 processor and 256 GB RAM. It is compatible with any system running MATLAB(R) and Parallel Computing Toolbox, including Windows(R), Linux, or Mac OS. GPU is not required for running this code.