GreenVision applies Near InfraRed Spectroscopy (NIRS) as a relatively new technology for the control of product quality and production processes. This technique can measure contents in a non-destructive manner, or analyse features in liquid, viscous and solid substances. NIRS has been intensively applied in the agro- and food industry in the past forty years. Although it had taken some time before the NIRS technique has been accepted, new developments in instrumentation and multivariate data analyis (chemometrics) has given NIRS the chance to grow into an invaluable analyis method.
NIRS involves measurement of the intensity of the near-infrared light absorbed by a sample as a function of its wavelength. The infrared light encompasses the wavelength area between 780 - 2500 nanometres (12,8000 - 4,000 cm-1) of the electromagnetic spectrum. The energy in this area corresponds with the energy of vibrations of molecular combination. In this context, NIRS is often also known as vibration spectrocopy. NIRS is used specifically for measurements of organic combinations (C-H, O-H, N-H en C=O). Multivariate data analysis is a must for the analysis of near-infrared spectra. The strong overlap among the absorption waves, caused by overtones and combination bands of stretching and bending vibrations, hinders the interpretation of NIR-spectra.
Since the extent of absorption in near-infrared is much lower than in the mid-infrared wavelength, NIRS can be used to measure concentrations to the order of a few to tens of percentages! In addition, excellent reflection spectra can be included. Glass fibre techniques can also be applied. All this opens possibilities for fast and non-destructive analysis, concurrently or otherwise, in various processes.
GreenVision FocusNIRS is an indirect measuring technique. This means that specific calibration models have to be calculated beforehand. Such calculations and the corresponding interpretation require professional knowhow and hands-on experience with these techniques. GreenVision has, over the years, much to do with NIRS, food research, and with application of multivariate regression techniques such as MLR (Multiple Linear Regression), PLS (Partial Least Square regression) and PCR (Principal Component Regression). Figure 4 shows the modelling process being tested in two steps. Firstly, the NIR spectrum and the reference measurement are used to calculate the unknown model. Subsequently, the calculated model is used, together with NIRS, to predict the value of the reference measurement.
In the area of Near Infrared Spectroscopy (NIRS), GreenVision works closely with the National Institute for Food Safety (Rikilt). This cooperation has widened the research arena, as much in the area of NIRS as in the area of research into agro- and food products. GreenVision and Rikilt are therefore the institutes to which queries can be addressed.
The making of NIR callibration models.
Commissioned by quality control laboratories, product associations and meat processing organisations, GreenVision has constructed different callibration models. The majority of the measured components are moisture, fats, proteins, carbohydrates, raw cell materials and certain sugars in different agriculture and food products. A few examples are: determining the composition of moisture, fats, proteins and ash in meat products, the texture of legumes, the composition of sugars in different kinds of (tropical) fruits, and the composition of oils in oil-rich seeds.
Developing a NIR measurement system for on-line moisture measurement.
Jointly with the parent organisation of the potato-processing industry, GreenVision is currently working on a NIR measurement system in which the moisture content in potatoes can be determined non-destructively on-line. The system has been constructed and is being fine-tuned.
- To determine the quantity of constituents in agro- and food products;
- To determine physical-chemical and sensory parameters;
- For qualifying raw materials, for example, for the pharmaceutical and food ingredients industry;
- Process control and optimisation
- Fundamental research: spectroscopic interpretation (vibrational and rotational bridging) and chemometry (experimental design, data pre-processing, callibration, classification, wavelength selection and validation);