All contained in these pages is (C) TOELT llc, and the name “Colour of Taste Technology” trademark has been submitted and is pending approval.
Recent advances in artificial intelligence have enabled to relax many hardware requirements exploiting the ability of the neural network to learn and compensate for the device-specific limitations. The implication is that analysis, which up to now were only possible with expensive instruments in a laboratory can be potentially be made accessible to a larger user base.
This technology allows us to build a compact and low-cost sensor that can learn to analyze substances based on fluorescence spectroscopy and using machine learning. The instrument consists of a compact and low-cost spectrometer controlled by a Raspberry-Pi and will use artificial intelligence to classify substances without using analytical substance-specific mathematical models.
This technology allows us to build a sensor based on a technology usable by people without technical training that is low-cost, portable (around 300 gr.), that can be used in any environment (on-site) and that can change the problem it solves on-siteUmberto Michelucci
The measuring principle of the instrument is fluorescence spectroscopy. Fluorescence is observed when substance, excited with light, absorb it and then re-emit light with lower energy. The light which can be absorbed, and the spectral properties of the re-emitted light are very specific of the substance and can be used to uniquely distinguish the substances. Among the beneﬁts of ﬂuorescence spectroscopy, are the high selectivity and sensitivity, and the non-contact and non-destructive nature of the measuring technique.
Example Use case: Quality control of oil
Due to the above-mentioned advantages, this technology is particularly suited for quality control and counterfeiting detection in the food industry. This approach has been proved to be able to distinguish, for example, different types of edible oils or detect the addition of adulterants to them.