Transmission Line Foreign Object Detection
Budget: $10 – $30 USD
I’m looking to train an object-detection model that automatically scans images or live drone footage of overhead transmission lines and flags anything that should not be there. The focus is foreign substances—plastic sheeting, stray metal parts, loose tools, or any other material that has attached itself to the conductors, insulators, or towers.
Here is what I need from you:
• A complete, reproducible training pipeline using
- GFRF R-CNN for high-accuracy detection of small foreign objects.
- YOLOv10 for real-time detection and deployment.
- Keras .
I will provide the Dataset.
Acceptance criteria: the model should reach at least 0.90 mAP on the validation set and run in real time, name of Foreign Object should be shown.
Here is what I need from you:
• A complete, reproducible training pipeline using
- GFRF R-CNN for high-accuracy detection of small foreign objects.
- YOLOv10 for real-time detection and deployment.
- Keras .
I will provide the Dataset.
Acceptance criteria: the model should reach at least 0.90 mAP on the validation set and run in real time, name of Foreign Object should be shown.