Overview:
This project focuses on detecting and counting cars moving in different directions (north, south, east, west) using computer vision. The YOLOv8l model is used to process images of size 640×640, with the ability to detect up to 10 different classes of vehicles, including cars and trucks. Vehicle movements are tracked based on lines drawn for each direction.
Key Features:
1. YOLOv8l Vehicle Detection:
The system uses the YOLOv8l model, which is designed for real-time object detection, to identify vehicles in images. It supports 10 different vehicle classes, including cars, trucks, buses, and motorcycles, providing flexibility in tracking various vehicle types.
2. Direction-based Tracking:
The system tracks vehicles moving in four directions (north, south, east, and west). For each detected vehicle, the model assigns a tracking ID and records the direction of movement based on lines drawn on the image.
3. Real-Time Counting:
Once vehicles are detected, the system automatically counts the number of vehicles traveling in each direction. This feature enables monitoring of traffic flow and congestion analysis.
4. Performance Optimizations:
YOLOv8l operates on 640×640 image resolutions, balancing accuracy and performance. This optimization ensures the system is efficient for real-time tracking even with high traffic volume.
Category:
Tags:
Links
Portfolio 
Leave a comment