I just thought I would post this as a discussion. I might look into it more when my vacation kicks in.
I recently took my car to the garage for its service, and that’s when I was informed one of the rear brake tail lights was out. I rarely use the car (I cycle further than I drive), its mostly the wife using it and I’m not proactively checking it.
So I was thinking, I have a security camera in the garage…I wonder if I can train something to recognize its the car (and not people) coming/going, take an image or two, and determine that the car has a problem with the any of the tail lights…brakes would come on and it sees one is ‘dark’? Something like deepstack or tensorflow based or is there a simpler way?