People Detection and Counting System

A real-time web-based platform that detects and counts people in images, videos, and live camera streams using the latest deep learning YOLOv8 architecture with GPU-accelerated OpenCV pipelines.
It is designed for corporate and individual use, enabling occupancy analytics, crowd monitoring, and automation triggers for smart environments.

---

Key Features

• Real-Time Detection: Achieves <200ms inference latency even on standard hardware.
• Multiple Input Modes:
  • Live detection from USB webcams, laptop cameras, and RTSP IP cameras (up to 30 FPS)
  • Upload photos (.jpeg, .png) and videos (.mp4, .mov) up to 1GB
• User Roles:
  • Trial User: Limited daily uploads (1 photo & 1 video per day)
  • Premium User: Unlimited uploads and real-time detection (monthly subscription)
• Output Data: Annotated frames and JSON-based summaries for seamless dashboard and mobile integration.
• Use Cases:
  • Library occupancy rate monitoring
  • Subway and public transport crowd analytics
  • Public restroom, hall, and event space usage tracking
  • Smart building automation triggers (e.g., emergency lighting, HVAC)

---

Technology Stack

• Backend: Python (Flask), Flask-Login, Flask-Mail, Werkzeug
• Frontend: HTML, CSS, JavaScript
• Detection Framework: YOLOv8 (Ultralytics), OpenCV
• Database: PostgreSQL, DBeaver
• Infrastructure: Docker, DigitalOcean (Production)
• Version Control: Git & GitHub

---

Workflow Highlights

• User Management: Registration, login, and profile editing with premium request handling.
• Media Upload & Processing: Users can upload photos or videos; detection pipeline returns annotated frames and counts.
• Live Detection: Supports both USB and IP cameras with real-time inference and visual feedback.
• Report Generation: Users can view and export occupancy analytics.
• Security: Email-based authentication and session management.

---

Example Applications

• Real-time monitoring of subway stations and libraries
• Automated reporting of restroom occupancy
• Fast, low-cost crowd analytics for events and emergency planning
• Drone-based aerial counting for industrial and agricultural use cases

---

Algorithms (Simplified)

• User Authentication: Secure registration and login system with email and password validation.
• Detection Pipeline: YOLOv8 for human detection, OpenCV for video feed handling, and optimized data pipelines for low latency.
• User Limits & Premium Mode: Daily upload limits for trial users; unlimited access for premium subscribers.
• Report Generation: Aggregated occupancy data and visualizations.

---

Getting Started

1. Clone this repository:

   git clone https://github.com/<siraeroglu>/people-detection-counting-system.git
   

2. Install dependencies:

   pip install -r requirements.txt
   

3. Run the application:

   python app.py