snt-ubix/escdd

E-SCDD Dataset

Overview

E-SCDD is an extended electroluminescence (EL) imagery dataset for photovoltaic (PV) module analysis.
It expands the existing BenchmarkELimages dataset by adding additional EL samples, refined pixel-level annotations, and standardized splits to support research on:

• defect detection
• anomaly segmentation
• robustness evaluation
• generative modeling
• PV modules' health monitoring and maintenance

Dataset Structure

• Images: EL imagery of PV modules, including multiple defect types such as cracks and inactive regions.

• Annotations: pixel-level segmentation masks provided in indexed PNG format.

• Splits: train, validation, and test folders following the original schema

Citation

This dataset extends and builds upon the BenchmarkELimages dataset, originally released under the MIT License.

Please cite both works when using this dataset.

@misc{hanifi2025bridgingdatagapspatially,
      title={Bridging the Data Gap: Spatially Conditioned Diffusion Model for Anomaly Generation in Photovoltaic Electroluminescence Images}, 
      author={Shiva Hanifi and Sasan Jafarnejad and Marc Köntges and Andrej Wentnagel and Andreas Kokkas and Raphael Frank},
      year={2025},
      eprint={2511.09604},
      archivePrefix={arXiv},
      primaryClass={eess.IV},
      url={https://arxiv.org/abs/2511.09604}, 
}

@article{pratt2023benchmark,
      title={A benchmark dataset for defect detection and classification in electroluminescence images of PV modules using semantic segmentation},
      author={Pratt, Lawrence and Mattheus, Jana and Klein, Richard},
      journal={Systems and Soft Computing},
      pages={200048},
      year={2023},
      publisher={Elsevier} }