Sunday, February 13, 2011

Superlative inflation?

Back when I was smaller and lighter than I am now, if I thought something was good I’d say, “That’s good,” and if it was really very good I might be tempted to say, “It’s great.” Today though, as my kids like to tell me, those phrases imply that something is not very good at all. What I should be saying is, “fantastic,” “amazing,” and “awesome,” preferably prefixed with the word, “totally,” as in ‘totally awesome.”

Okay, I’m old and grouchy. What of it?

Well I keep seeing references in the imaging industry press to “hyperspectral imaging.” At first I thought this was just one up on “megaspectral imaging,” or perhaps a big advance on “superspectral imaging,” but now I’ve learnt I was wrong.

Hyperspectral imaging, or so I’ve discovered, is the business of deducing chemical composition by imaging. I learnt this by reading a fascinating “Application Note,” “
Hyperspectral Imaging For Pharmaceutical Manufacturing Applications” on the web site of Headwall Photonics. Headwall, it seems are heavily into the business of building cameras that can do this kind of chemical analysis, and offer a whole range of hyperspectral imaging sensors. (Take a look at the movie on their web site, “In-Line Poultry Inspection - Hyperspectral Video” while you’re there.)

It’s a very interesting field – perhaps even totally amazing - and I can’t help thinking there are lots of latent opportunities for the technology.

1 comment:

Igor said...

The difference between multispectral and hyperspectral imaging revolves around the number of wavelength bands that can be acquired by the camera. Most cameras are RGB which is a rough three spectral camera although the red, green and blue are not "pure" spectral bands. In these multipecteal and hyperspectral camera, light is either divided in tens or hundreds (respectively) wavelength bands. The ability to have more bands eventually translate into the ability to do a better job a differentiating between similar looking objects. If the spectral bandwidth is narrow, then one can also do near spectrometric measurements, whereby one can make a difference between different chemicals.

One of the reasons you have not heard much about these cameras comes from the fact that they cost a pretty penny. With compressive sensing technology, those costs are likely to drop by one to two orders of magnitude:

http://nuit-blanche.blogspot.com/2011/02/dip.html


By the way, nice blog you have.

Cheers,

Igor.