---
license: cc-by-4.0
---

# Dataset Card for Electrical-Lines-Defect-Detection

- This repository contains datasets curated for APEPDCL's Line Quality Monitoring System, and will followed with further updates to datasets.
- These are released to foster open-source collaboration, reproducibility, and hands-on learning in AI.
- GitHub Repo Link: [Click here](https://github.com/EPDCL/Electrical-Lines-Defect-Detection)
---

## Overview
 
- **Captured by:** APEPDCL linemen with mobile cameras  
- **Curated by:** [Sampath Balaji](https://github.com/sampath-balaji/) & team, at [APEPDCL](https://www.apeasternpower.com/)
- **License:** CC BY 4.0
- **This ongoing project is supervised and mentored by [Sasank Chilamkurthy](https://chsasank.com/about/), whose expertise has guided its development.** 

---

## 📂 Repository Structure

```bash
Electrical-Lines-Defect-Detection/
├── Poles_LeanedStraight/
│   ├── ObjectDetection/    ← YOLOv12 bbox dataset + code    
│   └── Classification/     ← Folder + CSV classification dataset + code
└── README.md               ← (this file)
```

## Datasets
| Module               | Type           | Split            | HF Link                                                                                                                                                                                                                                                   |
| -------------------- | -------------- | ---------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Object Detection** | YOLOv12-format | train/valid/test | [Direct link to Dataset](https://huggingface.co/datasets/EPDCL/Electrical-Lines-Defect-Detection/tree/main/Poles_LeanedStraight/ObjectDetection) |
| **Classification**   | Folder+CSV     | train/val/test   | [Direct link to Dataset](https://huggingface.co/datasets/EPDCL/Electrical-Lines-Defect-Detection/tree/main/Poles_LeanedStraight/Classification)   |

## Usage
Each sub-folder contains its own ```README.md``` with:
- Dataset details (structure, labeling process)
- Training & inference instructions
- Metrics & sample outputs
Navigate into the folder you’re interested in, then follow the steps there.

## 🤝 Contributing
We welcome:
- Improved annotations or protocols
- Pull requests to code or documentation

## Sample Visualizations
- Classification model
<p align="center"> <img src="https://raw.githubusercontent.com/sampath-balaji/Electrical-Lines-Defect-Detection/refs/heads/main/Pole_LeanedStraight_Defect/Classification/assets/val_viz_ep40.png" width="600"/> </p>

- Object Detection model
  <p align="center">
  <img src="https://raw.githubusercontent.com/sampath-balaji/Electrical-Lines-Defect-Detection/refs/heads/main/Pole_LeanedStraight_Defect/ObjectDetection/assets/output.jpeg" width="600"/>
</p>

## 📄 License

- [Dataset](https://huggingface.co/datasets/EPDCL/Electrical-Lines-Defect-Detection): CC BY 4.0
- [Code](https://github.com/EPDCL/Electrical-Lines-Defect-Detection): MIT License