What all features does the node-red-contrib-opencv node support.?

General
1

I am intending to use NR for an academic IoT project. We have to build a drowsiness detection system that uses native opencv package. So I wanted to ask it in the community if using the node-red-contrib-opencv is the correct approach ?

I plan to implement the following features

  1. webcam integration
  2. real time processing of the webcam feed.
  3. Detecting eye contours in the video
  4. Alerts when the eye aperture is below a certain threshold.

Please provide some inputs if the node-red-contrib-opencv node can support these features.
I couldn't locate any solid tutorials on node-red-contrib-opencv, hence thought of collecting information to back my decision as I have tight deadline to deliver the project.

I am also facing issues in the installation of the node-red-contrib-opencv node.
If anyone can point out any foolproof installation guide, it would be a great help. I might create another thread to point out the actual issue in the installation.

Thanks a lot, appreciate it.

-Rajendra

2

Hi @rajendra-sahu,
Long time ago I have been added as a contributor for that node, but unfortunately I have never got time for it...
That node isn't maintained for more than 4 years, which is never a good sign. And if you look at the code, you will see that it only supports some classifiers. So not enough for you.

  1. Install opencv
  2. Grab a picture and send it to your server-side flow. E.g. using node/node-red-node-ui-webcam
  3. Write your own node - based on the code of the node-red-contrib-opencv node - to detect eye contours, and send those polygons in an output message.
  4. The next node in the flow will detect whether the polygon coordinates are below some level.
  5. The next node triggers some alarm.

There are lots of tutorials on the web about installing opencv. Don't focus too much on the readme page of that node...

Good luck with your assignment! You will have most work yourself on point 3...

Bart