Raspberry Pi is at the heart of this AI–powered, automatic sorting device that is capable of recognising and sorting any LEGO brick.
And its maker Daniel West believes it to be the 1st of its form in the world!
This mega-device was two several years in the building and is a LEGO creation alone, designed from about 10,000 LEGO bricks.
It can form any LEGO brick you location in its enter bucket into one particular of eighteen output buckets, at the level of one particular brick each two seconds.
When Daniel was motivated by earlier LEGO sorters, his creation is a huge move up from them: it can recognise completely each LEGO brick ever established, even bricks it has under no circumstances viewed ahead of. Consequently the ‘universal’ in the name ‘common LEGO sorting machine’.
The artificial intelligence algorithm behind the LEGO sorting is a convolutional neural network, the go-to for graphic classification.
What makes Daniel’s challenge a ‘world first’ is that he trained his classifier using 3D product visuals of LEGO bricks, which is how the device can classify completely any LEGO brick it’s confronted with, even if it has under no circumstances viewed it in serious everyday living ahead of.
Daniel has produced a full additional video (above) conveying how the AI in this challenge performs. He shouts out all the open resource program he utilized to run the Raspberry Pi Camera Module and entry 3D schooling visuals etc. at this level in the video.
LEGO brick separation
The vibration plate in motion, feeding solitary sections into the scanner
Daniel essential the enter bucket to carefully decide out a solitary LEGO brick from the mass he chucks in at when.
This is reached with a main and secondary belt slowly but surely pushing sections onto a vibration plate. The vibration plate makes use of a super fast LEGO motor to shake the bricks all around so they are not sitting down on best of each and every other when they get to the scanner.
Scanning and sorting
A Raspberry Pi Camera Module captures video of each and every brick, which Raspberry Pi 3 Model B+ then procedures and wirelessly sends to a much more effective computer system in a position to run the neural network that classifies the sections.
The classification choice is then despatched again to the sorting device so it can spit the brick, using a sequence of servo-controlled gates, into the appropriate output bucket.
Daniel is such a boss maker that he wrote not one particular, but two further looking at content for those of you who want to deep-dive into this mega LEGO creation:
Resource: Raspberry Pi website, by Ashley Whittaker.