Object Detection and Tracking with YOLO

This Python script utilizes the YOLO (You Only Look Once) object detection algorithm to detect and track objects in a video feed. It offers options for real-time preview, object tracking, and exporting detected objects.

Features

• Object Detection: Detects various objects in frames using YOLO models.
• Object Tracking: Enables object tracking for detected objects across frames.
• Export: Option to export cropped images of detected objects.
• Real-time Preview: Display real-time frame previews while processing.

Requirements

• Python 3.x
• OpenCV (cv2)
• Ultralytics
• Windows OS (for msvcrt usage)

Installation

1. Clone this repository:

   git clone https://github.com/iegrsy/YOLOv8_Test.git

2. Install required packages:

   pip install -r requirements.txt

Usage

When running the application, you can use the following parameters:

• -s, --video_source: File name or path of the video source. Default value: aoe.mp4.
• -e, --export: Option to export the processed video (True/False). Default value: False.
• -p, --preview: Option to enable preview mode (True/False). Default value: False.
• --skip_frame_count: Number of frames to skip. Default value: 1.

Example Usage

An example command to run the application:

python object_detector.py -s video.mp4 -p True -e True --skip_frame_count 2

This command runs the script on the 'aoe.mp4' video, enabling both object export and real-time preview.

Example training

Create dataset folder

cd build
mkdir datasets
cp -r person datasets/train
cp -r person datasets/test
cp -r person datasets/valid

Run training command

yolo task=classify mode=train data=datasets model=yolov8m-cls.pt epochs=2

License

This project is licensed under the MIT License.