NEC's winebot detects fraudulent bottles
By Evan Blass 
posted Aug 1st 2006 7:16AM
posted Aug 1st 2006 7:16AM
Apparently wine fraud is big business, with certain less-than-reputable vineyards watering down their Cabernet with a little Merlot, for instance, or even filling their supposedly boutique bottles with some off-the-shelf Manischewitz. Well the engineers / wine connoisseurs over at NEC have had enough, so they've invented a wine-tasting "robot" that can distinguish between 30 distinct varieties using only a five-milliliter sample. We've already seen NEC use infrared light in a food-tasting bot, and now they've taken a similar approach with this new device -- about twice the size of a three-liter wine box (mmm...box o' wine) -- which uses LEDs to fire infrared light at a sample and an array of photodiodes to interpret the resulting reflection. Until they can boost its recognition abilities to include many more varieties of wine (about 1,000, in fact), though, this project is little more than an academic exercise -- meaning you'll still need to bring your corkscrew to the liquor store if you want to ensure that you're getting a quality product.
Whenever I shop I do like to buy better brands but how do you always know that you are getting a quality product...this is a good step forward for wine.
I suppose if you can't taste the difference then why should you care?
I for one welcome our tipsy wine tasting overlords.
This might keep them at bay a little longer...
err, at least until the next morning when they wake with a bad brown-out and a killer CPU-ache! It could get ugly.
I, for one, welcome our robitic sommelier overlords.
Is this "robot" anything more than an infrared spectrometer? What exactly did these NEC guys "invent"?
I suppose if the technology got more sophisticated the government would be able to regulate wine producers to ensure that they are not selling fraudulent products to consumers. Although if someone manages to blindly choose a wine and not notice that it is a dodgy concoction of other substances rather then a true product, they should probably be sticking with beer.
by law, if a bottle has 75% of a certain amount of grape, then they are allowed to put that on the label.
eg. if it has 80% merlot, and 20% cabernet, then they are allowed to bottle it as merlot.
this is something everyone does, good or bad wineries.
what is the real purpose of this robot then?
That's why I drink Cain Cuvee - it has a blend of 5 different reds, as intended.
Instead of concentrating on a good "vintage" year, they blend their wines to have the same great taste every year.
http://www.cainfive.com/
And my favourite part is the photo of kit wine, which you of course would want to verify the authenticity of...
Dan2 is partially right. This is an application of an infrared spectrometer. What they 'invented' is the database search to make sense of the data the spectrometer collects. That's a non-trivial task. I've used IR spectrometers for many years... some of the database search 'matches' to my "unknown samples" are quite humorous. So being able to distinguish wine varietals is an accomplishment. Still, I think it's gonna be a few years before the robot kid has enough smarts for me. Meanwhile, I'm gonna stick with my sommelier to tell me if what's in the bottle is to my likin
Scientists just utilise the FTIR spectrometry (I suppose NIR region measurement) and everything else is only the statistics. If the amount of sugar is in the range of Cabernet wine, then also next 5 parameters must not exceed (free SO2, acids, alcohol, extract,...). Otherwise the Cabernet is not the cabernet. Theory is so simply. But they were FIRST. My respect. (They must to collect a lot of samples and also wines). I would like to work in, or just to clean up, that laboratory ;-).
Nowbody write that, this spectrometer is available only for White wine stored in bottles from clear glass! The IR spectrometry has a variety limitations and one of them is that the measured light must going through. I know, somebody can say, that we´ve got DRIFT or ATR techniques, but they penetrate to the sample only for few micro meters (by this techniques could be analyzed the glass of the bottle only, not the wine)... Don´t you think?