Rishit-dagli/People-Counter-On-Edge
📦 Open Source ProjectRishit-dagli
Low-latency, resilient people detection and counting solution optimized for edge computing devices.
People-Counter-On-Edge provides a streamlined architecture for deploying computer vision tasks on edge devices. By utilizing the OpenVINO toolkit, the project optimizes deep learning models to run efficiently on Intel-based hardware, ensuring low-latency inference. The system architecture includes a modular design: an FFmpeg-based server handles video stream ingestion, while the core detection engine performs real-time analysis to track and count individuals. Communication is managed via the MQTT protocol, allowing for lightweight, asynchronous data transmission to dashboards or other IoT infrastructure. This project is particularly useful for scenarios requiring privacy-preserving, local-only processing, such as retail analytics, occupancy monitoring, or smart building management. Its focus on resilience ensures that the counting process remains stable even in fluctuating network conditions or varying lighting environments.
💡Highlights
- ├─Optimized for Intel OpenVINO
- ├─Low-latency MQTT integration
- └─Resilient edge-based detection
🎯For
- ├─Edge AI Developers
- └─IoT Engineers