Hyperspectral_Imaging

Code for this paper.

Joint Spatial-Spectral Pattern Optimization and Image Reconstruction for Hyperspectral Imaging (IEEE-JSTSP 2022)

Tao Zhang, Zhiyuan Liang and Ying Fu.

Introduction

In this paper, we propose a deep-learning based method for high quality hyperspectral imaging by joint spatial-spectral optimization, including joint multispectral filter array (MSFA) and spectral sensitivity function (SSF) optimization, joint spatial demosaicing (SpaDM) and spectral super-resolution (SpeSR), and joint pattern optimization and HSI reconstruction.

Prerequisites

• Python >= 3.6, PyTorch >= 1.7.1

Getting Strated

1. Install the enviromemts

conda install -c conda-forge python-lmdb
conda install caffe
pip install --upgrade git+https://github.com/pytorch/tnt.git@master

pip install -r requirements.txt

2. Prepare datasets

• ICVL
• Harvard

To generate the taining and testing data, run python utility/lmdb_data.py

3. Trian

Spatial Optimization

python hsi_dm.py -a admmn_16channel_alpha -d /Path/to/Your/Data

Spatial-Spectral Optimization

• Selection

  python hsi_dm.py -a admmn_alpha -d /Path/to/Your/Data

• Design

  python hsi_dm.py -a admmn_spectrum -d /Path/to/Your/Data

4. Test

python hsi_test.py -a admmn_16channel_alpha --use-2dconv -r -rp /Path/to/Your/Model

Citation

If you find this work useful for your research, please cite:

@ARTICLE{zhang2022joint,
  author={Zhang, Tao and Liang, Zhiyuan and Fu, Ying},
  journal={IEEE Journal of Selected Topics in Signal Processing}, 
  title={Joint Spatial-Spectral Pattern Optimization and Hyperspectral Image Reconstruction}, 
  year={2022},
  volume={16},
  number={4},
  pages={636-648},
}