sumantadasdotdelhi/pfm1-kriging-hsi

PFM-1 Landmine VNIR Hyperspectral Imaging Dataset (IGARSS 2026)

Dataset Description

This dataset 1 contains Visible and Near-Infrared (VNIR) Hyperspectral Imaging (HSI) data prepared for the IEEE International Geoscience and Remote Sensing Symposium (IGARSS) 2026.

This specific version is a refined subset of the original benchmark dataset 2. While the original release provided full radiance cubes, broad GCP/AeroPoint data, and reference ground spectra of all the targets, this version focuses on a spatially subsetted region containing only PFM-1 landmine targets and includes high-accuracy, pixel-wise binary ground truth masks.

The data was collected over a controlled test field seeded with 143 realistic surrogate landmine and UXO targets (surface, partially buried, and fully buried). Data acquisition was performed using a Headwall Nano-Hyperspec® sensor mounted on a multi-sensor UAV platform flown at an altitude of ~20.6 m 2.

For more details regarding data acquistion and preprocessing, go to the original paper 2.

• Sensor: [Headwall Nano-Hyperspec®]
• Spectral Range: [398–1002 nm]
• Number of Bands: [270 bands]
• Approximate GSD: [Approx. 1.29 cm]

Dataset File Structure

The dataset is organized as follows:

File Name	Size	Description
site_with_only_mines	6.42 GB	Main Hyperspectral (HSI) data cube (ENVI format), referred to as "Full Region" in the paper [1].
site_with_only_mines.hdr	23.2 kB	Header file containing metadata for the HSI data cube.
roi_site_with_only_mines.roi	256 B	Region of Interest (ROI) file used for spatial subsetting the original dataset from [2].
binary_ground_truth_mask_for_pfm1	5.9 MB	Pixel-wise binary ground truth mask for PFM-1 mine locations.
binary_ground_truth_mask_for_pfm1.hdr	803 B	Header file for the binary ground truth mask.
target_signature_pfm_1_svc.txt	7.86 kB	Reference ground spectral signature for PFM-1 (captured via SVC).
.gitattributes	2.63 kB	Configuration file for Git LFS and XetData tracking.

Usage

Download

pip install huggingface_hub

from huggingface_hub import snapshot_download
 
snapshot_download(repo_id="SagarLekhak/pfm1-landmine-uav-vnir-hsi-IGARSS-2026", repo_type="dataset", local_dir="./data")

To load the data using Python

import spectral.io.envi as envi

# Load the hyperspectral cube
img = envi.open('site_with_only_mines.hdr', 'site_with_only_mines')
print(f"Loaded cube with {img.shape[2]} spectral bands.")

Citation

If you use this dataset, please cite the specific IGARSS 2026 work 1 and the original benchmark paper 2:

[1] IGARSS 2026 Work:

@misc{lekhak2026benchmarkingdeeplearningstatistical,
      title={Benchmarking Deep Learning and Statistical Target Detection Methods for PFM-1 Landmine Detection in UAV Hyperspectral Imagery}, 
      author={Sagar Lekhak and Prasanna Reddy Pulakurthi and Ramesh Bhatta and Emmett J. Ientilucci},
      year={2026},
      eprint={2602.10434},
      archivePrefix={arXiv},
      primaryClass={eess.IV},
      url={https://arxiv.org/abs/2602.10434}, 
}

[2] Original Benchmark Dataset:

@misc{lekhak2026uavbasedvnirhyperspectralbenchmark,
      title={A UAV-Based VNIR Hyperspectral Benchmark Dataset for Landmine and UXO Detection}, 
      author={Sagar Lekhak and Emmett J. Ientilucci and Jasper Baur and Susmita Ghosh},
      year={2026},
      eprint={2510.02700},
      archivePrefix={arXiv},
      primaryClass={eess.IV},
      url={https://arxiv.org/abs/2510.02700}, 
}