People Counter Docker
A distributed, microservice-based system for counting people in images using computer vision. A scalable, containerized solution leveraging message queues and AI models, designed to asynchronously process high volumes of visual data.
Screenshots
Technologies Used
Python 
OpenCV Architectural Context
The project is a fully distributed computer vision system designed with a microservice architecture. Its primary objective is the precise counting of people in images, maintaining high performance and reliability. Instead of a monolithic structure, I implemented a clean separation of concerns with independent services responsible for analytics and data storage.
Technology & Processing
The analytical core of the system is an advanced processing pipeline based on the YOLOv8 model, wrapped in high-performance FastAPI services. To ensure smooth operation under heavy load, I integrated RabbitMQ as a message broker. It manages an asynchronous task queue between the API and background workers, seamlessly forwarding the results to a dedicated database service.
Scalability & Deployment
The entire environment is fully containerized using Docker. A properly orchestrated compose configuration allows for the instant deployment of a cluster featuring multiple concurrent processing instances. This makes the system highly resilient, elastic, and ready to handle fluctuating traffic in surveillance systems, public space analytics, or retail monitoring.
Let's Build
Something.
Have a vision in mind? Let's engineer it together. Drop me a message and I'll get back to you as soon as possible.
