Raspberry Pi, Python and OpenGL

Hello Forums,

please take a look at the following youtube post. The user detects circles with from a webcam image in real time using openGL.


This is exactly what I need. It doesnt have to be realtime though. For this, consider the following image:


I need to detect the circular rims of the three cuvettes.

A few questions to start with:
Can this be done with openGL and Python on a raspberry Pi? Or is it better to use OpenCV? What is the difference between opencv and opengl? Both can do hough transform. One is relying purely on CPU, the other one is relying purely on GPU?

EDIT: In the opening credits of the video, the maker is saying: “desired shape is loaded and trained”… What does he or she mean by “trained”? I think Hough transform has nothing to do with neural networks

kind regards from germany,


I’d recommend using OpenCV, it has an OpenCL implementation and is better to use than implementing computer vision algorithms yourself.