YOLOv5 DeepSort Tracking Vehicle Network for Traffic Control

Introduction

🚀This repository introduces real-time tracking vehicle network for the traffic control system using YOLOv5 DeepSORT. The network is composed of two subsystems: edge-device tracking system and edge server. Edge-device tracking system used YOLOv5 DeepSORT, a lightweight machine learning model, as the major computing algorithm. After the edge-device tracking activates, it will send the detection result to the server to perform further analysis. For simulating network operation, we use a Node.js HTTP Module webserver as the communication protocol, and path generation as the main task of edge server.

Overview of source files

• app.js: Webserver host.
• track.py: Running (edge-device) vehicle detection and tracking system, sending data to the server.
• monitor_realtime.py: Running (edge-server) perspective transformation and visualization.

Installation

Require Node.js and Python to run.

For Node.js environments:

npm install express

For Python environments:

pip install -r requirements.txt