نظام ذكي لعدّ الأجسام باستخدام خط افتراضي
Smart Line-Based Flow Counter for Real-Time Object Detection
A YOLO-based object tracking and directional counting system that monitors objects crossing a specific line segment in real-time video. Designed for accurate flow analysis in controlled environments such as doors, exits, conveyor belts, and fenced areas.
Project Overview
This project implements a smart real-time counting system using the YOLOv8 deep learning model to detect, track, and count objects as they cross a predefined virtual line segment. Unlike traditional counters, it only triggers counting when objects pass through a specific portion of the frame, ensuring higher accuracy and eliminating noise.
Key Features
Custom Line-Based Counting
- 📏 Counts objects only when crossing a defined line segment.
- 🎯 Limits detection to specific X-axis boundaries.
Persistent Object Tracking
- 🆔 Assigns a unique ID to each detected object.
- 🔁 Prevents double-counting using ID history.
Direction-Aware Logic
- ⬇️ Counts only top-to-bottom movement.
- 🚫 Ignores reverse or invalid crossings.
Real-Time Visual Feedback
- 🎨 Bounding boxes, centroids, and counting line.
- 📊 Live display of total count on video.
Multi-Class Support
- 🐑 Tested on sheep counting.
- 🧍♂️ Works with people, vehicles, animals, and more.
Technology Stack
- 🐍 Programming Language: Python 3.x
- 🧠 Model: YOLOv8 (Custom Trained)
- 🎥 Computer Vision: OpenCV
- 🎨 Visualization: CVZone
- 📦 Libraries: ultralytics, NumPy, os
Use Cases
- 🐑 Sheep Exit Counting in farms.
- 🧍♂️ People flow monitoring at doors and exits.
- 🛍️ Retail customer exit analytics.
- 🚧 Construction site entry control.
- 🛤️ Conveyor belt item counting.
Conclusion
This smart directional flow counter provides a powerful real-time solution for object counting in constrained environments. By combining YOLOv8 detection, ID tracking, and spatial filtering, the system delivers accurate, scalable, and production-ready flow analysis for agriculture, retail, logistics, and security applications.