Data of the merged two public isolator detasets available with the code to merge then.

If you just want to use the data generated by this work, download the images and labels (in COCO format) available here.

All code was tested using Python 3.7

We use the Albumentations API to perform data augmentation.
The API is available here. Use the Albumentation instruction to install then.

Use the package manager pip to install the requirements.

pip install -r requirements.txt

To do the merge and increase the data of this work, it is necessary to download the file downsized.zip of the dataset proposed by Tomaszewski et al. available here and copy the content to the Tomaszewski folder. Do the same for the CPLID dataset available here and copying its contents to the CPLID folder.

Copy the contents of the 'albumentations' folder to the folder where the Albumentations project was cloned.

Change the "data_root" file path from data_augmentation.py file to the path that you cloned this project and run with Python in Albumentation project folder:

python data_augmentation.py

If you use this project in your research, please cite this project.

@misc{vieira2020unifying,
    author = {Vieira-e-Silva, Andr{\'e} L. and Chaves, Thiago and Felix, Heitor and Mac{\^e}do, David and Sim{\~o}es, Francisco and Gama-Neto, Milton and Teichrieb, Veronica and Zanchettin, Cleber},
    title = {{Unifying Public Datasets for Insulator Detection and Fault Classification in Electrical Power Lines}},
    year = {2020},
    howpublished = {\url{https://github.com/heitorcfelix/public-insulator-datasets}}
}

MIT