For this week Azalea and I trained a KNN classifier to differentiate different types of resistors. We used a webcam microscope to capture images of resistors placed on a piece of paper as background. We trained the model with about 200 images of each type of resistor.
The results are shown in the video below. This model doesn’t seem to be the optimal approach to differentiate resistors. Perhaps a better approach would be to use this to differentiate between resistors and other electronic components and then read the pixel color values to decode the resistor value without machine learning.
Classifying resistors using KNN and MobileNET from tinkrmind on Vimeo.