Other Example Applications
There are a wide range of potential applications for the Sagitto system. We've created the following demonstration models to give a small taste of how our miniature near-infrared spectroscopy system can be used to either quantify or classify materials. in just a few seconds.
A sophisticated NIR spectrometer : very small, portable, yet every bit as powerful as bench-top alternatives.
New Zealand hops, ripening in sunny Nelson
The female flowers - or 'cones' - on hop vines are rich in terpenoid essential oils and terpenophenolic resins. These 'alpha acids' give hops their characteristic bitter flavour. Hops with a higher percentage alpha acid content will contribute more bitterness than a lower alpha acid hop when using the same amount of hops.
The total alpha acid percentage varies between varieties and within specific varieties (depending on factors such as seasonal growing conditions, drying methods, and the point at which the hops are harvested.)
We refer to this model as our 'Scarborough Fair' model - it classifies dried herb samples into parsley, sage, rosemary, and thyme (and a few others ;-)
This model classifies ground coffee beans according to the coffee roaster, and the batch. Its not a very useful model in itself, but it does illustrate how near infrared spectroscopy can be used by more technically sophisticated coffee roasters to ensure that they produce an objectively consistent product.
Many specialised coffee roasters change their blends according to availability of particular beans. For example, the graph at left shows how the 'Rocket Espresso' house blend at Rocket Coffee changed when they altered the mix of beans in the 31 October roast, and then again when the blend was changed for the 4 November roast.
This demonstrates a classification model for plastic polymers, using near-infrared (NIR) spectroscopy data. As you can see from the graph at right, different types of plastic polymers have quite different absorption spectra in the 900nm - 1700nm wavelength range. Sagitto is able to use these differences to build very accurate classification models, using sophisticated machine learning techniques.
|Recycling Number||Abbreviation||Polymer Name|