🔍 Defect Analysis

Image-Based Solar Panel Defect Classification

Demo Mode
Simulated Classification Workflow

📋 About This Feature

This page demonstrates the defect classification workflow for solar panels. In a production system, this would integrate with a trained CNN model to analyze panel images and detect defects such as:

  • 🧹 Dust/Dirt Accumulation - Reduces efficiency by 15-25%
  • ⚡ Physical Damage - Cracks, scratches, or broken cells
  • ✓ Clean Panels - No visible defects
  • 🌳 Shading Issues - Blocked by vegetation or debris

Current Implementation: This demo simulates the classification process using the available image dataset (180 clean, 187 dusty images).

Access via API: Use /api/defects/classify/{anomaly_id} to get classification for any detected anomaly.

367
Total Images
180 Clean + 187 Dusty
✓
Demo Workflow
Ready for Testing
3
API Endpoints
Classify, Summary, Images

🔄 Classification Workflow

⚠️
1. Anomaly Detected
Power generation drop identified
→
📸
2. Image Analysis
CNN model processes panel image
→
🔍
3. Classification
Defect type identified with confidence
→
🛠️
4. Recommendations
Maintenance actions suggested

📚 API Documentation

GET /api/defects/classify/{anomaly_id}
Classify defect for a specific anomaly. Returns defect type, confidence score, and maintenance recommendations.
GET /api/defects/summary
Get summary of all defect classifications with counts by type and average confidence.
GET /api/defects/images
Get list of available clean and dusty panel images in the system.
📖 View Full API Documentation
Dashboard Anomaly Detection Defect Analysis