[katelynsills]

I work for a mill that cleans and sorts grains and beans (taking the rocks out, stems out, etc.), and it's fascinating to see the parallel invention of something really similar! We have a bunch of different steps:

1) Air is blown through the product and any dust is taken out. 2) The product is run through a bunch of screens that take out anything too big or too small. 3) The product is put through a gravity separator to separate based on mass. 4) Finally, the product is put through an optical sorter (https://www.youtube.com/watch?v=O0gWUeqzk_o) which uses blasts of air to push out unwanted materials from a stream of falling product.

I'm sure you could use the same process for Legos. Not sure about how to distinguish between branded and unbranded Legos though.

[jacquesm]

(4), That's a very neat machine!

What is the %age by weight of 'trash' versus 'good stuff' for such a sorter?

I do use screens for various pre-sorting stages, not shown in the article. The sorter is only good for parts up to 40 mm and anything that isn't a wheel or round so it will roll away while being imaged.

That's by far the bulk though so for me if it does that part well it is already more than worth it.

Branded/unbranded: spectrum is different (far more different than you would say by looking at it with the naked eye), weight does not match for the part (though this can be very close with really good fakes), logo on the studs is different.

I've been thinking about doing that gravity thing, but a bit more fancy, rather than just a binary sort to shoot parts in several directions, an alternative is a spiral slide under a steep angle where parts are fed in at the top and ejected when they reach the right bin.

That's a lot more complicated to make than what I have right now mechanically and also the time available for a classification operation would be much shorter, but it would allow for a much larger number of output bins without taking up a whole lot of space. So maybe a next generation, if I still need it (this one is going through piles of lego now).

[unityByFreedom]

Cool project. How many bins are you organizing your Lego into? Or was this more just a proof of concept thing?

Plans to launch a Lego sorting service? ;-)

[jacquesm]

> How many bins are you organizing your Lego into?

Seven, so it takes multiple passes before it is done.

> Or was this more just a proof of concept thing?

Tough question :) No, it's for real it really has to sort through the 2000 kg, but if it needs to be beefed up or changed to get to the end then I'll do it. The next step 'up' would be a machine designed from scratch incorporating all the lessons learned with essentially unchanged software. There are still some limitations that could be addressed but then you'd lose the training set and you'd have to start all over from scratch. That might be worth it to get the last 1% error or so, so if this ever becomes 'real' then I'd have to do that. I highly doubt it will get to that though. time will tell.

> Plans to launch a Lego sorting service? ;-)

Not at present, though you're not the first person to think of that, parents with kids are suggesting I should make it mobile to visit people at their homes for $x / shot :) Still, that will only happen if I really have nothing better to do, which means likely never.

[unityByFreedom]

> Still, that will only happen if I really have nothing better to do, which means likely never.

Don't sell yourself short. You built the thing, after all. Could be some fun road trips with, I'm sure, gracious hosts to entertain you during Lego sort. That could be a whole retirement life right there ! ;-)

[ekux44]

I'd pay 3-4 figures for an on-site sort of my personal Lego collection.

[jacquesm]

Oh if you put it that way :)

[keithpeter]

Remember to charge by the 100Kg sorted rather than the days on site :-)

[danielharan]

You can probably use your already sorted legos to re-create a training set with a new machine.

[chillydawg]

Problem with that is the "errors" in the real output bins will be encoded into the next generation of neural net and no progress will be made. On the other hand, the dataset could be used as a starting point that could be one-time filtered and improved by a human.

[Mtinie]

Congrats, @unityByFreedom, I think you're the first person to suggest "LSaaS".

The VCs should start lining up in ...five...four...three...

:)

[unityByFreedom]

Well, it wouldn't surprise me.

OP is familiar with the secondary marketplaces and knows if he could classify and sort well enough to make sets, then could potentially make money buying bulk and reselling sets on these marketplaces.

It could be a real business. He's already proven he has the chops to design it. If he doesn't, I don't doubt someone else will try.

Obviously Lego itself could do this but probably make more money from melting and recycling their own parts under their brand as new, like Apple does with its recyclers. Do we know if Lego is recyclable as new Lego by melting?

[jacquesm]

No, Lego does not recycle. But there is a very lively secondary market.

There was a short period that Lego allowed people to design their own sets and order them through Lego but it was so popular they had to shut it down.

People even buy new sets to sort them out just to get new bricks to combine into their own creations without having to buy them in bulk and not being able to use half of what they buy.

[unityByFreedom]

Well then, it sounds like Lego isn't interested in furthering either a custom or secondary market which could undercut their new sales. That is a business opportunity, no?

Could you make money just buying bulk, sorting out rare parts, and reselling those?

[drostie]

It seems like the wheel problem could be solved by a dowel rod that would keep things rolling along until they fell over? I suppose that unfortunately gives another way for the thing to get clogged up?

I'm surprised that this strategy isn't more quicksort-ish, I was expecting a first pass into warm or cool colors, then another pass into large or small lego, etc.

[jacquesm]

Everybody that hasn't sorted lego in bulk says the same thing :) The color sort would not help because in the end the idea is to get the same shaped parts in to a bin, not the same colored parts. Our brains are much better at picking a red part out of a background of similar shaped parts than they are at picking out a red part of a certain shape from a background of other red parts with different shapes.

As for the multiple passes, that's actually how it works, the first pass sorts into the 7 most common categories and then quickly works down from there. But the 'runoff' bin is by far the largest at the end of every run. After 3 passes the bulk has completely disappeared and most of the remains are sufficiently rare to be dumped into the right container individually. It goes pretty quickly: first pass: 7 bins, second pass: 7 bins again, but 49 if you put the first seven through again (but in larger volume accumulated from several runs on smaller lots). And so on. Since the bulk of the lego is in those first 49 and the value is in the remainder it doesn't take long to add value to a pile.

As for the mechanics, any obstruction, no matter how well intended will immediately become a point for a bridge to anchor to.

[spullara]

What is the back of the enveloper calculation of doing it by hand? Doesn't seem like it would take a few high school students that long to go through the 2000kg and come out with a very nicely sorted set of Lego bricks.

[jacquesm]

About 4 man years based on what I've done myself.

[bigger_cheese]

I work in a steel mill we use a similar setup with an optical camera to size streams of Ore and Coke particles. One thing you could look into is using a vibrating feeder (sometimes called a 'vibro') this is what we use to stop screens from 'pegging' - similar issue to the bridging problem mentioned in the article.

[Baeocystin]

High-speed optical sorters are the sort of behind-the-scenes tech that make me feel like I'm living in the future. I have vivid memories as a child watching people winnow rice by hand using wide, flat bamboo baskets.

Had I seen something like the optical sorter back then, I would have thought it (Arthur C. Clarkeian) magic!

[spyder]

Yep, they are very fast:

https://www.youtube.com/watch?v=YvU2KEl5_ZE

https://youtu.be/vbSww5SBqN4?t=75

https://youtu.be/zYimxuEAEtk?t=67

[scoot]

> it's fascinating to see the parallel invention of something really similar!

This isn't parallel invention - the principal of optical sorting and air ejection are well known and understood. (Which is not to lessen the achievement in building this, but building and inventing are not the same thing.)

[nommm-nommm]

Dunno who here is old enough to remember but back in the day every bag of chips used to contain a few burnt chips. Well, thanks to those computer vision air blasting sorting machines theres no more burnt ones in the bag! They all get air blasted out and now chips are uniform.

[microcolonel]

Sorting beans was the exact thing I was thinking of when watching this. I process a lot of chickpeas at home, and the quality of the dried chickpeas going in is quite mixed. There are some split ones (not good for sprouting), and some duds which won't soak or sprout even though they're whole.

I think that people might pay for the convenience of getting a bag full of proper chickpeas with no duds and no stones (I've heard of this, but never really seen it).

Maybe some day Suraj and NuPak will go a little further with their cleaning and get rid of damaged beans and not just stones.

Though, frankly I think they could do better than average with a specific gravity/densimetric sorter; it would probably require fewer passes through the optical sorter as well.

[pbhjpbhj]

Sorting beans is a well-solved problem using a sort of waterfall (of beans) where computer vision selects the bad beans and air jets eject them.

The equipment costs money; and reduces the volume of "product" left for sale. Your problem is capitalism - the companies don't do that because it lowers their profit.