Adding a simple Ansible playbook and some updated notes/requirements, for easy deployment of monitor_run.py

[wombat70]

I've written a basic Ansible playbook to automate (and document) some of the setup/configuration steps required to get the monitor_run.py script running on a Raspberry Pi with the provided yolov5 model. There is also a short README which outlines how to run the playbook.

For those who prefer to configure things manually, the playbook YAML file together with the revised requirements.txt file effectively documents the steps/tasks and required dependencies. This may also help resolve/workaround issue #1.

All the new files are in a new deploy directory and I've only made minor changes to the monitor.py script, namely:

• Use filetype instead of imghdr, because the latter is being deprecated ref. PEP 594.
• Added some OpenCV settings in an attempt to reduce capture latency when using a USB camera.
• Use results.ims instead of results.imgs, because the yolov5 Detections class renamed its image-list property in Improved Profile() inference timing ultralytics/yolov5#9024.

Hope this helps! I'm looking forward to receiving any comments/suggestions or questions. :-)

Tom

[spienz]

Hy,

works so far with Rasbian 64, but:
SSH Key Config should be getting a better description.

And with ansible the installation of opencv won't work. We had to install it manually with pip "opencv-python-headless"
Then it works fine so far.
We should implement a test picture function.
And also document the way to geat an IP Camerea working. We solved it for us.

Regards

Walter & Sebastian

[andrw3000]

The three new additions are extremely useful, thanks for the contribution. Amendments to monitor_run.py lgtm.