For many kids, crusts are the bane of a delicious PB&J sandwich, but they eventually grow up to embrace a bread’s chewy outer wrapping. There are some holdouts, however, such as Andrew DeGonge who continues to despise crusts so much into adulthood that they built a smart automated robot that can automatically de-crust a sandwich.
Slicing up a sandwich isn’t an especially difficult task, but there’s always a risk when handling knives, so the safer alternative is to put a robot in charge of the knife instead. What could possibly go wrong? DeGonge’s machine is assembled from an aluminum frame with custom 3D-printed parts that are all moved around by belts and stepper motors to ensure a high level of precision because you want to sacrifice as little sandwich as possible.
A Raspberry pi 4 Model B paired with a Raspberry Pi Camera Module V2 take advantage of OpenCV to automatically photograph and analyze any sandwich dropped onto the robot’s cutting board, determining its shape, calculating the largest rectangle to create, and then dropping the knife four times like a guillotine while spinning the sandwich on a turntable to remove the crust on every side.
It’s a surprisingly capable solution, but there are some ways it can be improved. For starters, watching the video of the robot in action it becomes clear the knife DeGonge is using isn’t sharp enough to lop the crusts clean off. It’s in desperate need of an upgrade, or a good sharpening. The approach the robot uses also seems optimized for sandwiches made of very square bread, not rounder loaves which are, arguably, more delicious. The four perpendicular slices it makes can dramatically decrease the size of some sandwiches as a result, so a more nuanced approach to the computer vision analysis and how the bot sliced and dices would be a welcome upgrade.