gayatrigovindasetty/vehicle-detection

🚗 Real-Time Vehicle Detection System

This space demonstrates real-time vehicle detection using YOLOv8.

Features

• —Detect 8 classes of vehicles: Auto, Bus, Car, LCV, Motorcycle, Multiaxle, Tractor, Truck
• —Process images and videos in real-time
• —Adjustable confidence and IOU thresholds
• —Powered by YOLOv8 Nano (~50 FPS on GPU)

Usage

1. 1.Image Detection: Upload an image and click "Detect Vehicles"
2. 2.Video Detection: Upload a video and click "Process Video"
3. 3.Adjust detection parameters as needed

Model

The trained model is automatically downloaded from Hugging Face Hub at startup.

Performance

• —Speed: ~50 FPS on GPU, ~10 FPS on CPU
• —Accuracy: mAP@0.5 > 85%
• —Model Size: ~6 MB

Built with Ultralytics YOLOv8, Gradio, and Docker.

Source Code

GitHub Repository