Detection of Empty Shelf using YOLOv9

Empty Shelf Detection using YOLO & Roboflow

Out-of-stock shelves cost the retail industry billions every year. Manual shelf auditing is slow, inefficient, and unreliable — so we built an AI-powered solution to solve this problem.

What We Built

A real-time Empty Shelf Detection system that uses a custom-trained YOLO object detection model to automatically identify empty or out-of-stock spaces on retail shelves through a live mobile camera feed.

Tech Stack

  • YOLO — Real-time object detection
  • Roboflow — Dataset annotation, augmentation, and training
  • YOLOvX App — Live mobile deployment
  • Custom Dataset — Real retail shelf images manually annotated

How It Works

  1. Point the camera at a retail shelf
  2. The YOLO model processes the live feed
  3. Bounding boxes appear on empty shelf spaces
  4. Total empty space count is displayed instantly

Real-World Impact

  • Faster restocking decisions for store staff
  • Reduces revenue loss caused by out-of-stock situations
  • Eliminates the need for manual shelf auditing
  • Scalable for CCTV-based fully automated monitoring systems

What We Learned

This project gave us hands-on experience in:

We are proud of what we built and excited to continue exploring AI-powered retail automation solutions.