How ragelayasree-5725/infrasight works
InfraSight is a local road-damage inspection prototype: a user supplies a road image, the FastAPI application runs YOLO detection, and the result includes damage classes, confidence, and bounding-box evidence. The application also exposes inspection history, per-detection details, and metrics, with SQLite-backed detection records and locally served output images. The README documents these capabilities; the sampled Python excerpts show the API, inference, and storage mechanisms, but omit some wiring and the HTML contents, so uncertain links are left out.
Main entry points: main.py FastAPI application · index.html Image dashboard · camera.html Camera page · best.pt Trained YOLO model
Components 16
Inspection interface
- Image dashboard
index.html - Camera page
camera.html
Application API
- FastAPI application
main.py - Image upload
main.py - History and metrics
main.py - Detection details
main.py
Detection pipeline
- YOLO inference
main.py - Trained YOLO model
model/best.pt - Class and confidence
main.py - Severity assessment
main.py - Box visualization
main.py
Inspection records
- Detection images
main.py - Detection database
main.py - Record and history logic
main.py - Persistence and priority
main.py
Other components
- Inspector
Connections 18
- Inspector → Image dashboard: opens
- Inspector → Camera page: opens
- Image dashboard → Image upload: submits image
- Camera page → Image upload: submits image
- FastAPI application → Image upload: routes requests
- FastAPI application → History and metrics: routes requests
- FastAPI application → Detection details: routes requests
- Image upload → YOLO inference: runs detection
- YOLO inference → Trained YOLO model: uses weights
- YOLO inference → Class and confidence: produces detections
- Class and confidence → Severity assessment: assesses damage
- YOLO inference → Box visualization: plots boxes
- Box visualization → Detection images: writes image
- FastAPI application → Detection images: serves files
- Record and history logic → Detection database: reads and writes
- History and metrics → Record and history logic: loads history
- Detection details → Detection database: queries record
- Persistence and priority → Record and history logic: supports assessment