Smart Transportation Monitoring System

An intelligent traffic monitoring system that uses AI to detect and track vehicles in real time. It provides a live dashboard, REST/WebSocket APIs, and an AI chatbot for traffic queries.

Architecture Overview

Short Demo

Video.3.mp4

Features

• Real-time vehicle detection and tracking (YOLO + ByteTrack)
• Multi-camera processing with multiprocessing
• Live dashboard with traffic analytics
• REST + WebSocket APIs for realtime data
• WebRTC low-latency video streaming
• AI chatbot (LangGraph ReAct) for natural language queries
• Model optimization with INT8 OpenVINO and TensorRT
• Model pruning with torch-pruning (see prune.py and readme_torch_prunning.md)
• CPU and GPU support

Tech Stack

• Backend: FastAPI, Python
• Frontend: React, TypeScript, Vite
• Cache/Queue: Redis
• Object Storage: MinIO
• Database: PostgreSQL
• AI/ML: YOLO, ByteTrack, LangGraph

Requirements

• Python 3.11+
• Node.js 18+
• PostgreSQL 16+
• Redis
• MinIO
• gdown (for downloading sample videos)
• NVIDIA GPU (optional)

Setup

Manual

1. Copy environment files:

# macOS/Linux
cp backend/.env.example backend/.env
cp frontend/.env.example frontend/.env

# Windows PowerShell
Copy-Item backend/.env.example backend/.env
Copy-Item frontend/.env.example frontend/.env

2. Edit the .env files to match your local services and secrets:

• If Redis or MinIO are already running, update REDIS_URL and MINIO_* to match.
• If not installed, install and start Redis and MinIO, then update the values in backend/.env.

3. Download sample videos into backend/app/video_test:

cd backend/app
gdown --folder https://drive.google.com/drive/folders/1gkac5U5jEs174p7V7VC3rCmgvO_cVwxH

4. Install backend dependencies:

cd backend
pip install -r requirements_cpu.txt  # or requirements_gpu.txt

5. Install frontend dependencies:

cd frontend
npm install pnpm
pnpm install

6. Run the backend:

cd backend
uvicorn app.main:app --reload --host 0.0.0.0 --port 8000

7. Run the frontend:

cd frontend
pnpm run dev

Backend: http://localhost:8000 Frontend: http://localhost:5173

Docker

1. Copy environment files (same as manual).
2. Ensure backend/app/video_test contains the sample videos (use gdown if needed).
3. Start services:

docker compose up --build

GPU build (optional):

DEVICE=gpu docker compose up --build

Backend: http://localhost:8000 Frontend: http://localhost:5173

Main APIs (v1)

Base prefix: /api/v1

REST

Auth

• POST /auth/register - Create a new account
• POST /auth/login - Login and receive JWT
• GET /auth/me - Get current user profile

User

• PUT /user/password - Change password
• PUT /user/profile - Update profile info

Traffic

• GET /road/roads_name - List monitored roads
• GET /road/info/{road_name} - Current traffic stats (counts, speeds, status)
• GET /road/history/{road_name} - Traffic history (paginated)
• POST /road/webrtc/offer/{road_name} - WebRTC session setup (SDP offer -> answer)

Chat

• POST /chatbot/chat - Send a message to the AI assistant

Chat History

• GET /chat-history/messages - List chat messages
• POST /chat-history/messages - Save a message
• DELETE /chat-history/messages - Clear all messages
• DELETE /chat-history/messages/{message_id} - Delete a message
• GET /chat-history/messages/count - Count messages

Admin

• GET /admin/resources - System metrics (CPU, RAM, Disk, Network)
• GET /admin/traffic/status - Worker status per road
• POST /admin/traffic/roads/{road_name}/start - Start a road worker
• POST /admin/traffic/roads/{road_name}/stop - Stop a road worker

Realtime (WebSocket / WebRTC)

• WS /road/ws/frames/{road_name} - JPEG frame stream
• WS /road/ws/info/{road_name} - Realtime traffic metrics
• WS /road/ws/chart/{road_name} - Realtime chart data
• WS /chatbot/ws/chat - Realtime chat stream
• POST /road/webrtc/offer/{road_name} - WebRTC signaling for low-latency video

Authentication

Most endpoints require JWT. Admin endpoints require admin role.

Header:

Authorization: Bearer <TOKEN>

WebSocket query:

?token=<TOKEN>

API Docs

Swagger UI: http://localhost:8000/docs