Hyperspectral Image Classification via a Random Patches Network

Matlab implementation of our method for hyperspectral image classification.

Paper

Hyperspectral Image Classification via a Random Patches Network

Please cite our paper if you find it useful for your research.

@article{rpnet,
  title={Hyperspectral image classification via a random patches network},
  author={Xu, Yonghao and Du, Bo and Zhang, Fan and Zhang, Liangpei},
  journal={ISPRS Journal of Photogrammetry and Remote Sensing},
  volume={142},
  pages={344--357},
  year={2018}
}

Installation

• Clone this repo.

git clone https://github.com/YonghaoXu/RPNet

Dataset

• We have uploaded the Indian Pines image and the corresponding annotations into the dataset folder.

Usage

• Add the directory path for the dependencies files in RPNet.m.

addpath(genpath('/utils/'))
addpath(genpath('/dataset/'))

You may change these two commands according to your environment.

• Run RPNet.m.

Acknowledgement

• The LibSVM package from Prof. Chih-Jen Lin is utilized for the SVM classification in this repo.