How 4run4v/citytrack-ai works
CityTrack AI is intended as a city-scale, multi-camera vehicle and plate tracking system with traffic analytics and a command-center frontend. The documented domain stages include camera streams, vehicle detection and tracking, ANPR, and analytics; the repository also includes a FastAPI backend and PostgreSQL/PostGIS schema. Sampled implementation is mostly placeholder code, so component-to-component runtime wiring is not established here. The only concrete API behavior sampled is the backend health endpoint.
Main entry points: main.py FastAPI Backend
Components 18
Camera Ingestion
- Stream Manager
camera/camera_manager.py
Vision and ANPR
- Vehicle Detection
vision/detection/vehicle_detector.py - Vehicle Tracking
vision/tracking/tracker.py - Plate Detection
anpr/plate_detector.py - Image Preprocessing
anpr/preprocessing.py - Plate OCR
anpr/ocr.py - Plate Normalization
anpr/plate_validator.py - Local AI Models
models
Traffic Analytics
- Congestion
analytics/congestion.py - Traffic Density
analytics/density.py - OD Matrix
analytics/od_matrix.py - Speed Analytics
analytics/speed.py - Traffic Analytics
analytics/traffic.py
Application and Data
- Command Center
frontend - FastAPI Backend
backend/main.py - PostgreSQL/PostGIS
database/schema.sql
Other components
- IP Cameras
- HTTP Client
Connections 2
- HTTP Client → FastAPI Backend: requests health
- FastAPI Backend → HTTP Client: returns status