YOLOv8 Video Inference with OpenCV

This repository contains two Python scripts that leverage YOLOv8 for video-based object detection. The scripts allow you to perform inference either in real-time using a webcam or on a pre-recorded video.

Features

• Real-Time Inference: Use the webcam to process live video streams.
• Offline Inference: Analyze and annotate a pre-recorded video file.
• Performance Metrics: Measure and display the inference time and FPS.
• YOLOv8 Integration: Utilize the YOLOv8 model for robust object detection.

Prerequisites

Before running the scripts, ensure you have the following dependencies installed:

• Python 3.8 or higher
• OpenCV (pip install opencv-python)
• Ultralytics YOLO (pip install ultralytics)

Scripts Overview

1. VIDEO CAP.py

This script uses the webcam to perform real-time object detection.

How to Run:

python "VIDEO CAP.py"

2. COMPUTER VISION video.py

This script processes a pre-recorded video file and performs object detection on each frame.

How to Run:

1. Place your video file in the desired directory.
2. Update the video_path variable in the script to point to your video file.
3. Run the script:

   python "COMPUTER VISION video.py"

Configuration

Both scripts require a pre-trained YOLOv8 model. Update the model path in the following line:

model = YOLO("/path/to/your/yolov8-model.pt")

Real-Time Script (VIDEO CAP.py)

The camera index is set to 0 by default. Modify this line to use a different camera:

cap = cv2.VideoCapture(0)

Video Script (COMPUTER VISION video.py)

Replace video_path with the path to your video file:

video_path = "/path/to/your/video.mp4"

Output

• Annotated frames are displayed in a window named "YOLOv8 Inference".
• Press q to exit the program.

Example Outputs

• Real-Time Inference:
  • Frames are processed and displayed with bounding boxes and labels for detected objects.
• Video Inference:
  • Annotated video frames are displayed with detection results.

Troubleshooting

1. Webcam Issues:

   • Ensure no other application is using the webcam.
   • Try changing the camera index (0, 1, etc.).

2. Video File Not Opening:

   • Verify the video_path is correct and points to a valid video file.

3. Performance:

   • Use a system with a dedicated GPU for faster inference.
   • Reduce the video resolution for better FPS.

License

This project is licensed under the MIT License. See the LICENSE file for details.

Acknowledgments

• Ultralytics YOLO for the YOLOv8 framework.
• OpenCV for video handling and visualization.

Contributing

Feel free to open issues or submit pull requests to improve this project!