Hacker News new | past | comments | ask | show | jobs | submit login

A device seems like overkill. Write code to apply machine vision to this grid: https://possiblywrong.files.wordpress.com/2019/04/skittles_a...



I think the fastest and most accurate process would be to:

* empty a pack of skittles onto white paper and take a photo

* use some existing image recognition libraries to count each color

* export to csv or smn and do excel magic


Even without an image recognition library, it should be possible to simply count the pixels and how close to red, green, blue, etc. they are to know how many skittles there are in the picture.


You'd probably want a jig for that, wouldn't you?

Hence, building.




Consider applying for YC's Spring batch! Applications are open till Feb 11.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: