Dataset · EPDCL
Smartphone photos of power grid defects
EPDCL/pd-defect
Field images of electrical distribution infrastructure with annotations for detection, segmentation, and classification.
What's inside
Inside are 20,476 annotated photographs of electrical distribution network objects. The shots were taken in field conditions on smartphones in collaboration with the power company APEPDCL in Andhra Pradesh, India. The data includes annotations for object detection, image classification, keypoint detection, semantic segmentation, and perspective verification.
What you can do with it
- Training CV models for automatic damage detection on utility poles and power lines
- Creating mobile power infrastructure inspection systems using smartphone photos
- Fine-tuning segmentation and detection models for industrial and energy tasks
- Tasks
- Object detection, Image classification, Keypoint detection, Image segmentation
- Size
- 10K–100K rows
- Downloads this month
- 20.5K
- Likes
- 8
How to load it
from datasets import load_dataset
ds = load_dataset("EPDCL/pd-defect")