Home >> PRODUCTS >> NurScan – quality control with NIR sensors

Quality control for small businesses

Food authenticity

Imagine owning a small tool, smaller than a smartphone, capable of providing you with the information you need for that particular product: maturity, contamination, fat…This is the reason we want to make a system of quality control with NIR sensors.


NurScan is our application for agri-food production. It was developed following the SANUM research project.

We use the near infrared scanner (NIR) SCiO from Consumer Physics Inc. to obtain a “spectrum”. In fact we can consider each spectrum like a fingerprint of each product. Specific chemometric analysis make it possible to create “models”. We compare new samples of that particular kind of product to the model. Like this we obtain relevant information about their authenticity. Go to our Research and Development pages for more detailed information.

We offer customized “models” for each of your products. Your company can select the parameters to control. The goal is to offer you a quality control system with NIR sensors.

Use of SCiO sensor for cheese analysis

How does it work

Sample collection

As a producer you know which products and which parameters you are interested in keeping under control. We work together with you for the choice of the samples and during the data collection. We use the SCiO scanner to obtain the fingerprints.

The number of samples to collect depends on the product and the analysis you want to do.

Sample analysis

During the collection the spectra are checked visually. We perform various tests to check their quality and their specificity.

We have created a database to collect the data – spectra, sample properties, sampling parameters. Like this we always have all the information available. In particular it allows us to easily identify possible problems in sampling.

Databases and models

After we have analyzed the correct number of samples for a specific product we can start producing the so-called “models” using chemometrics and statistics algorithms. These algorithms group the spectra that we had obtained based on the chosen specific properties. We can create different models for the same samples of a product. It will depend on the type of verification we want to perform.

Validation and testing

Validation of models is an important step to understand their reliability. For this, we use tester samples whose properties have been ascertained through other analysis. The model is considered reliable when the results obtained from these tests are in line with the analysis. And of course with what we already know about them.

Use in your company

Once all the previous steps are completed, it is time to use NurScan in production thanks to an easy-to-use smartphone app, and a SCiO scanner! After a short training, you will be able to carry out quality control analysis in your company for all the selected steps of your process.

Contact us for more information