Analog Devices and Consumer Physics, the makers of the Scio pocket spectrometer, have formed some sort of partnership to develop the miniature NIR spectrometer into something you can embed in a mobile device. 

Scio says they're shipping dev kits now for the pocket version, which looks pretty cool.  The idea is that Consumer Physics maintains a large library of spectra on some cloud server and compares your spectra to that library.  Also, developers can build small, custom calibrations for particular sample sets.

There is come complaining that the spectrometer is a toy because it doesn't have the performance (mostly in spectral resolution) of a lab instrument and can't see trace quantities (they say the detection limit is around 1%).  But it seems like there are a lot of applications for an inexpensive consumer instrument that can tell you concentrations of major constituents, even with some limitations on sample size or homogeneity.  The app everybody seems to mention where you point it at your pork chop and it tells you it's a pork chop and how many calories pork chops have is kind of stupid, but maybe somebody will write an interesting food app for cooking or brewing or some other home chemistry.

I'm learning a lot about remote sensing and hyperspectral image processing, so I'm going to talk about that stuff here.  The plan is to learn the basics of those fields and write some introductory stuff based on that and also to try to follow news and interesting applications and post about those. 

To kick things off, here's a link to a special issue of the Lincoln Labs journal on spectral imaging for remote sensing.  It's from 2003, but it's a general enough treatment to be completely relevant.