How to do pattern recognition and counting?

Question

Machine learning

I need to count all the objects on the electrical circuit
. What is my problem ... I need to recognize the objects on the circuit, which are always the same, and have a total number, approximately like in the photo, have each individually and in common, and so on for each element (photo from the Internet), that is, to count individually or all, if possible, from a pdf file, or jpeg.
(in short, then have one image, and use it to count the rest of the similar ones in the diagram, have the exact number)
Help me find some way out, preferably not difficult, since I am a complete zero in this

Answer 1

https://docs.opencv.org/master/d4/dc6/tutorial_py_...

Answer 2

I would do something like this in matlab/skylab:
1. Select the green paint layer using color spaces and get a b/w mask.
2. Perhaps I would work with morphological filters, if necessary.
3. Found related bwlabel components and calculated their regionprops properties.
4. Selection of components by area, topology and/or other features invariant to rotations in accordance with templates.