Skip to content

Latest commit

 

History

History
52 lines (35 loc) · 1.75 KB

README.md

File metadata and controls

52 lines (35 loc) · 1.75 KB

Hyperspectral anomaly change detection based on autoencoder

Pytorch implementation of JSTARS paper "Hyperspectral anomaly change detection based on autoencoder". image

Paper

Hyperspectral anomaly change detection based on autoencoder

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

@ARTICLE{9380336, author={Hu, Meiqi and Wu, Chen and Zhang, Liangpei and Du, Bo}, journal={IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing}, title={Hyperspectral Anomaly Change Detection Based on Autoencoder}, year={2021}, volume={14}, number={}, pages={3750-3762}, doi={10.1109/JSTARS.2021.3066508}}

Installation

Install Pytorch 1.10.2 with Python 3.6

Dataset

Download the [dataset of Viareggio 2013] 链接:https://pan.baidu.com/s/1x_M0nRqV-jmugIB6MltmXQ 提取码:ogum

[Dataset]: "Viareggio 2013" with de-striping, noise-whitening and spectrally binning

img_data.mat:

img_1(D1F12H1); img_2(D1F12H2); img_3(D2F22H2)

pretrain_samples:

un_idx_train1,un_idx_valid1,un_idx_train2,un_idx_valid2; [acquired from the pre-detection result of USFA, Wu C, Zhang L, Du B. Hyperspectral anomaly change detection with slow feature analysis[J]. Neurocomputing, 2015, 151: 175-187.]

groundtruth_samples:

un_idx_train1,un_idx_valid1,un_idx_train2,un_idx_valid2;

random_samples: un_idx_train1,un_idx_valid1,un_idx_train2,un_idx_valid2;

Usage

maincode.py

More

🌷Homepage🌷
🔴Google web 🔴
🌏ResearchGate 🌍