yolov8_webcam

Repo to showcase using yolov8 with your webcam in python

cv2-yolov8.py

This Python script uses OpenCV and YOLO (You Only Look Once) for real-time object detection. The script captures video from the webcam, applies the YOLO model for object detection, and saves the output video with detected objects.

Dependencies

• OpenCV
• Ultralytics YOLO

pip install opencv-python ultralytics

Directory Structure

The script uses two directories:

• outputs: This directory is used to store the output videos. If the directory does not exist, the script will create it.
• models: This directory is used to store the YOLO models. If the directory does not exist, the script will create it.

Usage

You can specify the type of YOLO model to use with the -model_type argument. The options are:

• seg: for segmentation models
• pose: for pose estimation models
• obb: for oriented bounding box models
• cls: for classification models

If no model type is specified, the script will use the default YOLO model.

Example

python cv2-yolov8.py -model_type seg

Models

Models will automatically be downloaded if you do not have the specific model installed locally

Output

The script saves the output video in the outputs directory. The filename is in the format {model_type}_{timestamp}.mp4, where:

• model_type is the type of YOLO model used
• timestamp is the current date and time in the format YYYYMMDD_HHMMSS