- Live detect when dog pooped
- Play alert sound
- Send notification via Pushbullet
- Attach detection image in notification
- Deployable to Raspberry Pi
This project uses Poetry for package management. If you don't have Poetry installed, you can install it using the following command:
pip install poetry wheel
Once you have Poetry installed, you can install the project dependencies using the following steps:
- Clone the repository:
git clone https://github.com/dennislwy/dog-poop-detector-yolov5.git
- Navigate to the project directory:
cd dog-poop-detector-yolov5
- Install the project dependencies:
poetry config --local virtualenvs.in-project true
poetry install --only main -v
This will install all the dependencies listed in the pyproject.toml
file.
Now, you should be able to run the project. For example, to start live detection with an RTSP stream, you can use the following command:
python live.py --weights best.pt --view-img --nosave --notify-img --source rtsp://your_rtsp_url
RTSP Stream
python live.py --weights best.pt --view-img --nosave --notify-img --source rtsp://your_rtsp_url
Testing with MP4 video
python live.py --weights best.pt --view-img --nosave --no-notify --source dataset/tests/test1.mp4
yolov5 detect --conf-thres 0.7 --line-thickness 2 --view-img --weights best.pt --source dataset/tests/test1.mp4