Remote Sensing Stacked Autoencoder and Clustering framework for Geological Mapping

Accepted: Advances in Space Research

This repository provides code and supplementary materials for the paper entitled 'Remote sensing framework for lithological mapping via stacked autoencoders and clustering'. We present a framework based on different dimensionality reduction methods, including principal component analysis, canonical autoencoders, stacked autoencoders, and the k-means clustering algorithm to generate clustered maps using multispectral remote sensing data which are interpreted as lithological maps.

Abstract:

Supervised machine learning methods for geological mapping via remote sensing face limitations due to the scarcity of accurately labelled training data that can be addressed by unsupervised learning, such as dimensionality reduction and clustering. Dimensionality reduction methods have the potential to play a crucial role in improving the accuracy of geological maps. Although conventional dimensionality reduction methods may struggle with nonlinear data, unsupervised deep learning models such as autoencoders can model non-linear relationships. Stacked autoencoders feature multiple interconnected layers to capture hierarchical data representations useful for remote sensing data. This study presents an unsupervised machine learning-based framework for processing remote sensing data using stacked autoencoders for dimensionality reduction and k-means clustering for mapping geological units. We use Landsat 8, ASTER, and Sentinel-2 datasets to evaluate the framework for geological mapping of the Mutawintji region in Western New South Wales, Australia. We also compare stacked autoencoders with principal component analysis and canonical autoencoders. Our results reveal that the framework produces accurate and interpretable geological maps, efficiently discriminating rock units. We find that the accuracy of stacked autoencoders ranges from 86.6% to 90%, depending on the remote sensing data type, which is superior to their counterparts. We also find that the generated maps align with prior geological knowledge of the study area while providing novel insights into geological structures.

Code Requirements

• Linux and Windows are supported, but we recommend Linux for performance and compatibility reasons.
• 1+ high-end NVIDIA GPU for sampling and 1+ GPUs for training.
• 64-bit Python 3.9 and PyTorch 2.1 (or later). See https://pytorch.org for PyTorch install instructions.
• Other Python libraries: pip install click Pillow psutil requests scipy tqdm diffusers==0.26.3 accelerate==0.27.2

Preparing Dataset

This framework is applied to three different data types, including are in /datasets/main_dataset/*.zip folder or can be download from the source attached below.

A small dataset also included in the datasets/sample_dataset.zip folder.

1. Landsat-8 download here

2. ASTER download here

3. Sentinel-2 download here

Full dataset:

download here

Notebooks for each dataset: main run code.

• We share the code for each dataset separately including all the experiments.

• Note that the specific dataloader, data preprocessing and postprocessing should be done by users depending on particular datasets.

  Autoencoder_Landsat8.ipynb

  Autoencoder_ASTER.ipynb

  Autoencoder_Sentinel2.ipynb

Proposed method: flow overview.

Results Elbow plot for each dataset and methods.

License

Copyright © 2024,Transitional Artificial Intelligence Research Group & AFFILIATES. All rights reserved.

All material, including source code and pre-trained models, is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

References

  @article{NAGAR2024,
  title = {Remote sensing framework for geological mapping via stacked autoencoders and clustering},
  journal = {Advances in Space Research},
  year = {2024},
  issn = {0273-1177},
  doi = {https://doi.org/10.1016/j.asr.2024.09.013},
  url = {https://www.sciencedirect.com/science/article/pii/S0273117724009335},
  author = {Sandeep Nagar and Ehsan Farahbakhsh and Joseph Awange and Rohitash Chandra},
  keywords = {Remote sensing, Deep learning, Dimensionality reduction, Stacked autoencoders, -means clustering, Geological mapping},
  }

Sandeep Nagar, Ehsan Farahbakhsh, Joseph Awange, Rohitash Chandra, Remote sensing framework for geological mapping via stacked autoencoders and clustering, Advances in Space Research, 2024, , ISSN 0273-1177, https://doi.org/10.1016/j.asr.2024.09.013.