Semantic Segmentation for a Solar-Panel Inspection Drone
Overview
What this challenge is about.
Fine-tune a U-Net for pixel-level solar-panel segmentation on drone images, then evaluate IoU and deliver a field memo. Earn a verifiable certificate.
The scenario
The startup (around 45 staff) inspects around 80 rooftop arrays/month in Spain and Portugal; pixel-accurate dust-coverage estimates would let them quote precise cleaning contracts.
The Brief
What you'll do, and what you'll demonstrate.
Train a panel-segmentation model that achieves at least 0.80 IoU on the held-out test set and recommend the next data-collection priorities.
Earning criteria — what you'll demonstrate
- Fine-tune a small segmentation model on a real dataset
- Design augmentations that mimic the deployment distribution
- Evaluate segmentation with IoU and qualitative inspection
- Recommend data-collection priorities from error analysis
Program Fit
Where this fits in your program.
Sharpens the same skills your degree expects you to demonstrate.
Aligned coursework coming soon.
Skills
Skills you'll demonstrate.
Each one shows up on your verified credential.
- Semantic Segmentation
Apply semantic segmentation to solve real industry problems and demonstrate production-level capability.
- Cnn
Apply cnn to solve real industry problems and demonstrate production-level capability.
- Transfer Learning
Apply transfer learning to solve real industry problems and demonstrate production-level capability.
- Data Augmentation
Apply data augmentation to solve real industry problems and demonstrate production-level capability.
- Pytorch
Apply pytorch to solve real industry problems and demonstrate production-level capability.
- Evaluation
Apply evaluation to solve real industry problems and demonstrate production-level capability.
Careers
Career paths this challenge builds toward
Completing this challenge demonstrates skills that transfer directly to these roles:
Computer Vision Engineer
Domain-adapted segmentation on aerial imagery is the day-one CV engineering task at any drone-inspection or geospatial-AI company.
This challenge sharpens
- semantic-segmentation
- cnn
- transfer-learning
Machine Learning Engineer
Augmentation discipline and clear evaluation are the MLE habits that get domain-shift problems solved.
This challenge sharpens
- data-augmentation
- evaluation
- pytorch
Applied AI Scientist
Turning error analysis into next-campaign data-collection priorities is the applied-AI-scientist craft of closing the loop between model and data.
This challenge sharpens
- evaluation
- data-augmentation
- semantic-segmentation