Fruit-sorting robot ‘will disrupt the industry’

http://www.drivesncontrols.com/news/fullstory.php/aid/4938/Fruit-sorting_robot__91will_disrupt_the_industry_92.html

by Futurebot • 11 years ago

55 119 11 years ago

amelius2• 11 years ago

I'm wondering what is so difficult about this task. I mean, all the "good" fruit looks almost the same, so recognizing it should be simple. Also, there's plenty of training material.

sageabilly• 11 years ago

Fruit is not a uniform shape and it typically bruises easily so it can be difficult to design a robotic "hand" to consistently pick up fruit of irregular shape without bruising.

This article from Gizmodo does a good job of explaining the difficulties involved: http://www.gizmodo.com.au/2015/10/why-is-it-so-hard-for-robo...

amelius2• 11 years ago

Looking from those images, I think they are using the wrong kind of hand. A better hand would simply be a "vacuum cleaner" end, i.e., gripping the fruit by suction power. If it can't be gripped like that, it is probably not fruit that needs to be selected.

shas3• 11 years ago

I'm wondering what is so difficult about this task.

Well, conceptually it sounds simple. But it is certainly a complex engineering problem. You are forgetting building conforming actuators and sensors, controlling these, image registration, the low margin for error, classification performance, real-time processing, etc. etc.

Applying the principle of capitalism, if it was simple to design and if a market and technology enables it, then it should already exist. In this case, the market obviously exists (e.g. farm worker demand in fruit industry), so if the solution doesn't already exist, then it is highly likely to be because the problem isn't simple to solve and/or the technology is inadequate.

oldmanjay2• 11 years ago

Even more generally: assuming things that you haven't done must be simple is not a valid argument, just a silly opinion.

fnordfnordfnord2• 11 years ago

I used to develop food sorters. Color sorting is very simple in principle and much harder in practice. The main difficulties are good lighting, not shaking your cameras and optics to pieces and keeping the optics and lights clean.

For fruit sorting in particular there is the very difficult task of conveying the fruit without damaging it. The best sorter in the world won't help you if it mangles the fruit.

In my old lab, there are solutions to most of the common food sorting problems, not all of those solutions make it out to the field though.

scientist2• 11 years ago

How big is this market?

rubidium• 11 years ago

I toured a commercial apple orchard 10 years ago. They use massive parallel conveyor system with laser-imaging for quality and gates to sort.

I'd be surprised if this robotic arm+vision system could compete on cost and through-put.

Retric2• 11 years ago

Note this "can distinguish and selectively pick fruit." Right now you either set up nets, have a machine bulk shake a tree and take what falls, or pay someone to pick ripe fruit. The conveyor system would then sort the collected fruit.

This system either replaces human collectors or significantly increases yields compared to shaking the trees. But, you would likely still need the bulk sorting at a later stage.

sitkack• 11 years ago

I think this would free migrant farm labor from having the burden of income, freeing them to be wondering poets. Reminds me a recycling sorting robot, which in my country is done by the mentally disabled. Just the kind of person you want to take a job away from.

forrestthewoods2• 11 years ago

I agree. Innovation and technological advances cause more harm than good. Break the threshing machines! Break them all!!

xtian6• 11 years ago

The industrialization of agriculture has made clear inroads on the problem of world hunger and has markedly increased the quality of food from the time when all farming was manual labor.

joosters2• 11 years ago

Don't forget to ban the plough. These unemployment causing disruptive technologies are ruining farming!

wolfram742• 11 years ago

Jobs are pretty shitty, they get in the way doing any real work. If we can have a functional society where robots deal with all the manual labor, the owners of the robots get half the profits and the other half get spread uniformly amongst the populace, anybody who wants to work on things that demonstrably unprofitable like writing linux kernel, or teaching their children to read, or advanced math or analyzing adventure time for asinine thematic story telling, well, who knows what we might get done.

moron4hire• 11 years ago

Unfortunately, that's not what is going to happen. The owners will just take all of the profits, maybe even get themselves a subsidy passed to promote the use of robots under some "green energy initiative" or something.

hugh42• 11 years ago

And remind me why I, as a farmer (hypothetically) who spent millions of dollars buying a piece if land and millions more to buy robots to work on it, then spends all his time worrying about maintenance and shipping and prices and pruning and planting and weeds and so forth... remind me why I should consent to having my profits confiscated and given to other organisms for no reason other than that they happen to be of the same species?

Imagine the social unrest when a large portion of the populace is literally of negative value. It's not going to be pretty.

We need to start thinking now about how we're going to achieve a smaller but more elite population, so only useful people get born.

sinxoveretothex2• 11 years ago

Well, if your view on work is that we shouldn't make processes more efficient if it means taking jobs away, how about this: automate work and then pay people to dig a hole and fill it, paying them with the profits from the automation.

Now, maybe you think it'd be quite useless, so you might suggest cutting away the crap and just paying people to do nothing. Congrats, you are now a proponent of basic income!

klipt• 11 years ago

Do the robots get basic income too? I mean eventually they're gonna agitate for their rights...

josu• 11 years ago

This is a bit off topic, but this GIF of a carrot harvesting machine never ceases to amaze me: http://i.imgur.com/UkTalyS.gifv

Toast_• 11 years ago

Efficiency is a beautiful thing.

thomasahle• 11 years ago

But carrots are easier than fruit. If you go to apple or orange plantages, you'll often see the ground covered in perfectly good fruit, because it's too expensive to pick it efficiently. 20% doesn't even make it to the shop: http://www.forbes.com/sites/nadiaarumugam/2012/10/04/un-says...

In addition, being able to remove weeds mechanically, rather than chemically, is a big win for health and environment.

julianozen• 11 years ago

"Once picked, the fruit can also be sorted by colour so that, for example, red apples can be separated from green apples."

So it can compare apples and oranges. Thanks I'm done

FussyZeus• 11 years ago

You beat me to that joke. Well played.

wcummings• 11 years ago

I don't know, people are really really good at things like this (picking fruit), which robots are generally speaking very bad at, and migrant labor is cheap. Why is this a disruptor?

hugh42• 11 years ago

For starters, not every country is the US, we don't all have vast hordes of illegals willing to work for next to nothing.

Secondly, and without wanting to start a political thread, the US might not either soon because President Trump.

Overall there's a point at which this becomes economical (just as there was a point at which automated harvesting of every other crop became economical starting with the 19th century) and it depends on labour costs in each country. It won't be super-soon, but it will happen eventually.

Jack000• 11 years ago

Never thought I'd see someone endorse trump on hn. September never ends I suppose.

Shivetya• 11 years ago

not sure if I buy this, while the idea sounds cool the complexity of that gripper tells me that there have to be simpler means to grip fruit without damaging it that would be faster, something along the lines of suction mechanisms where you can stack many of them; like on a roller.

What I am getting at is, besides the obvious complexity of the "hand" is how many can you fit into the area and how fast can they move the product? Identifying the good product from bad isn't hard, many food production lines do that and blow off items they don't want.

compostor4216• 11 years ago

Automation like this is why I call BS on the "we need more immigrants to keep the economy going!" folks.

We will be struggling to keep our existing population employed. The last thing we want to do is import even more low-skilled labor who will soon be made redundant.

Increasing automation means a declining population is a godsend rather than a problem.

Policy makers must be aware of this. I suspect the call for increased immigration are politically motivated (replacing the existing population with a more reliable voting block) rather than sincere economic concerns.

This recent article is a great example. The author gleefully explains how replacing the existing white population with immigrants will reduce the NRA's base.

https://www.washingtonpost.com/posteverything/wp/2015/10/19/...

oldmanjay2• 11 years ago

The calls for restricted immigration are certainly politically motivated (mainly variations on "I don't want them getting mine") so what's the difference? National borders mean less and less all the time. Clinging to them to weather this upcoming economic storm is like using an anchor as a liferaft.

mason2403• 11 years ago

Almost no one in the US is calling for restrictions on immigration.

compostor4216• 11 years ago

So why not let the entire population of Africa move to the United States? Borders are an antiquated concept according to you.

BuckRogers5• 11 years ago

You had me agreeing until the part about the NRA base going down with increased immigration. Now I'm all for it.

compostor4216• 11 years ago

Purposeful demographic replacement of white people to achieve political ends. Lovely.

13thLetter3• 11 years ago

The government should dissolve the people and elect a new one, then? Even as a joke that's in very poor taste.

compostor4216• 11 years ago

So if black people in your state were being replaced, and that helped your pet political issue, you would cheer on their replacement?

Somehow I doubt you'd say that in polite company. Funny how blatantly racist rhetoric is okay... if the recipients are those mean old whites!

scrollaway3• 11 years ago

Automation like this is why I call BS on the "we need more immigrants to keep the economy going!" folks.

This is a bit of a shortcut (and most don't realize it) to "more people doesn't necessarily mean less jobs to go around, as available work is not a finite product because of the amount of artificial demand an increased population can create".

We don't need more human drones; when properly funneled, increased immigration can mean a better economy.

compostor4216• 11 years ago

We don't need more human drones; when properly funneled, increased immigration can mean a better economy.

Agreed. Importing millions of unskilled laborers is insane as automation picks up. Those jobs will be made redundant in short order. (They probably already would have been, but a massive pool of cheap labor meant little incentive for automation. Similar to the first steam-engines being unused by the ancient Greek ruling class http://www.mlahanas.de/Greeks/HeroAndLoon.htm)

Bringing in the select best and brightest from other countries is certainly wise though.

JoeAltmaier3• 11 years ago

I think those numbers don't add up. Decreasing population does nothing for the per capita need for working-class people. Unless you think laborers are dying off at a greater rate than the rest of us.

compostor4216• 11 years ago

I don't understand, can you explain?

Increased automation definitely decreases the per capita need for working-class people.

We would probably have a lot more automation already (especially in low-end agriculture/service-industry) if not for a seemingly inexhaustible supply of cheap labor from south of the border.

dalacv• 11 years ago

Import? We are not importing them. They migrate here because there is work. If there is no work, they won't come over here. It is basic supply and demand that is controlling this.

compostor4216• 11 years ago

From the perspective of the elites who encourage unchecked mass immigration to drive down labor costs and secure reliable voting blocks "import" is quite accurate.

acchow• 11 years ago

What about the immigrants in high-skilled labor?

The recent calls for increased immigration are mostly about STEM-trained immigrants, are they not?

compostor4216• 11 years ago

Vast majority of immigrants are unskilled.

Letting in a few of the best and brightest from other countries makes sense. Importing millions of central/south American peasants is idiocy.

cheald3• 11 years ago

The last thing we want to do is import even more low-skilled labor who will soon be made redundant.

This is the lump-of-labor fallacy. More immigrants means greater demand for labor (and other goods and services), not just greater supply and results in net job creation and wage increases for native workers. Automation isn't likely to change that, because the demand for labor is not fixed.

hchenji3• 11 years ago

Yes I agree...instead of importing low skilled labor, we should import highly skilled immigrants. Increasing the H1B and employment based greencard quotas is absolutely necessary.

supergirl• 11 years ago

DISRUPT

Animats2• 11 years ago

Fruit sorting systems have been in production use for years. They're very impressive. Vision systems, AI, robotics, and the ability to handle huge quantities of produce.

Here's a production scale apple sorting and packing system.[1] The part of the video for "big plant" shows the most automated version.

Here's a system for sorting peas.[2] Each individual pea is inspected using computer vision. Each individual pea. Anything that doesn't look like a perfect pea is kicked out by an air jet. Take a look at the speed at which the system is running, as a wide conveyor belt feeds hundreds of peas per second through the machine, with the good ones coming out on one conveyor and the rejects on another. That one machine is doing a job a thousand hand sorters could not do.

Automated cherry grading.[3] This video has a good explanation of what's going on, as size, color, stem length, and defects are all checked by a computer vision system. That system can process over 700Kg/hour of cherries. "We're enabling our clients to massively reduce their sorting staff."

Tomato sorting machine.[4] This video has slow-motion sections so you can see the tomatoes being sorted. The machine is far too fast for humans to even see what it's doing.

There are vision-based sorting machines for almost every kind of produce, and they're in wide use all over the world. All this automated sorting means that customers no longer need to examine produce in the supermarket. The rejects were removed at the plant, and turned into animal feed or something. This makes services like Amazon Fresh possible.

[1] https://www.youtube.com/watch?v=cprDuf0ASzU [2] https://www.youtube.com/watch?v=JyGR6A5MWG0 [3] https://www.youtube.com/watch?v=Y3jlq03Gviw [4] https://www.youtube.com/watch?v=Lz88nsWL4kw

aidenn02• 11 years ago

I think you underestimate the quality of a thousand hand sorters. That machine may replace on the order of a hundred hand sorters, but I doubt it's 1000.

Animats2• 11 years ago

6 metric tons of peas per hour per machine. The machine examines the pea from all sides while the pea is in free fall. A human would have to pick up each pea to do a comparable job. A kilogram of peas is about 4000 peas, so 1000 humans would each have to examine 24,000 peas per hour each, or 6 peas per second, to keep up.

Retric2• 11 years ago

Your confusing the sorting aspect with the 'picking' aspect. Right now a lot of fruit ends up on the ground and never makes it to the sorting facility.

scientist2• 11 years ago

The machines showed in these videos seem suited for very large and expensive processing facilities, which probably are cost-effective because one supplies the consumption of millions to tens of millions of people. But how smaller producers can automatize their processing without owning or depending on these facilities? Would robots that cost of the order of tens of thousands euros / dollars, that automate more flexibly but slowly than the machines in the videos, be cost-effective for smaller producers? How big is the market for such robots?

vegabook4• 11 years ago

Capital trumps labour, yet again. Talk about a brave new (terrible) world.

How is anybody without amazing talent, IQ, and luck in the birth-location lottery, supposed to survive?

How are these awful machines helping to advance the wellbeing of humanity? By destroying livelihoods for a few pennies cheaper bananas to those who can already afford to pay more?

sliverstorm2• 11 years ago

How are these awful machines helping to advance the wellbeing of humanity?

Hypothetically, by freeing humanity from hard manual labor and making goods less costly.

Machines have a long history of advancing the wellbeing of humanity. The fire bow, wheel, pottery wheel, bicycle, ox cart, lever, ramp, plow... I could go on and on, from ancient to modern.

The modern textile industry is a much closer advancement, that made clothing manyfold more affordable and freed countless people from unforgiving constant manual labor.

vegabook4• 11 years ago

That sorting fruit is "hard labour" is actually coincidental. The real skill is pattern recognition and dextrosity. Skills widely deployed by humans today. We're now on the cusp of destroying this, another whole class of human employability. I am not as confident as you that we can simply extrapolate the past ad infinitum (ie: the "we will adapt and flourish" dogma). It seems to me we're reaching some inflection points in the trend where real, painful adjustment will occur. Recall that the 19th century's "innovations" created vast misery then, and war for 50 years 50 years later, before we adjusted. For those of us (un)lucky enough to be alive today, the immediate future seems very bleak indeed. Thanks to thinking machines.

kmicklas2• 11 years ago

The suffering this thing would cause is a failure of our economic and political systems... not the fault of technology.

ItsDeathball• 11 years ago

In the long run, humanity is better off if we're not paying people below minimum wage under the table to pick pears. If a portion of the surplus value generated by these awful machines can be captured and paid to the people they displace, those people can then do something they actually enjoy.

vegabook4• 11 years ago

Very true if, and only if, "a portion of the surplus value generated by these awful machines can be captured and paid to the people they displace".

But a very small group of financiers/entrepreneurs/VCs own all the machines, their ownership of the machines gives them an advantage in owning even more machines, and it is not in their nature (ie, human nature) for them to spread the sunshine. That's the problem. That's why we have a vast underclass already today, and it's not getting any smaller, despite the wishful thinking of the technerati (of which I am one).

vacri• 11 years ago

If we get enough of this kind of technology, then Basic Income might become a reality, through sheer political necessity.

ArtDev• 11 years ago

Cool. Robots. http://www.fanuc.eu/uk/en/robots "With more than 100 models, FANUC offers the widest range of industrial robots in the world."

birbal2• 11 years ago

I have some personal experience with applying computer/machine vision techniques to assessing rice quality. In India, rice quality assessment is a high value activity as the difference in per unit weight price is dramatic across different varieties of rice. The longer the grain, as an example, the pricier it is. Discoloration is also an indicator of bad quality rice. Since there is such a huge difference in price, there is naturally cheating, subjectivity and sampling driven errors. We had developed a device that would objectively and quickly measure the rice quality based on a multitude of factors. This was not just software because there had to be way to cancel out any variance due to ambient conditions and it had to be a device that was easy to maintain in a highly rugged rice factory environment. We did really well but where we failed was our inability to break through the powerful rice quality inspectors cartel and even the rice suppliers who were worried about being "found out". Manual inspection was strangely seen as way of having some control over the proceedings.

anonbanker• 11 years ago

As an ag/tech startup, I'm very interested in this device. is there a way to contact you for more information?

birbal2• 11 years ago

This was 7 years back my friend. We built software prototype for measuring dimensions, classifying grain type and measuring color / brokenness. For hardware we had some initial designs where the objectives were - to easily put grains on a platform so they are separated from each other, easy cleaning and lighting uniformity. We had conceived some rotating or vibrating disks for grain separation. But the mechanical part of the hardware wasn't actually built. I have moved on. There is nostalgia around that was triggered by this thread, but nothing more at this point. Sorry to disappoint.

Fruit-sorting robot ‘will disrupt the industry’

Related Stories