The Challenge
Retail stores and shopping centers need accurate visitor counts to optimize staffing, evaluate campaigns, and understand customer flows.
Traditional counting systems are expensive, inaccurate, or require complex infrastructure.
The Solution
We developed an AI-based counting system running on Raspberry Pi with YOLOv8 computer vision.
The edge device detects people crossing a defined line and sends data via MQTT to a FastAPI backend. A React dashboard shows real-time data, historical trends, and analytics with interactive charts. The system is GDPR-compliant — no images are stored, only anonymized count data. It runs stably at 5 FPS with accurate detection.
The Result
People Counter delivers reliable real-time visitor counts at a fraction of the cost of commercial systems. Stores gain insights into customer flows that enable them to optimize staffing and measure the impact of campaigns and store layouts.