Software solutions for NIR handheld spectrometer



Software solutions for NIR handheld spectrometer

The software module offers three different views on the data:

  • The basic visualization of NIR raw spectra
  • The difference view
  • The cluster view

Most powerfull is the automatic identification of known substances.


The visualisation of the NIR raw data

It shows the measured spectrum with a resolution of 10 nm in the range of 900 nm to 1,700 nm. To manually compare plastics with already stored materials, stored spectra can be loaded and displayed.

A very straightforward, but powerfull approach for evaluation  of NIR raw spectra.

Visualization of NIR raw spectra with 10 nm spectral resolution.

How does the difference view work?

In the presentation of the raw spectra no significant differences can be seen.

In the difference display, the different raw spectra are subtracted from each other (metaphorically speaking). Ranges in which the spectra are identical show “zero” difference. Ranges in which the NIR spectra are not identical show differences.

The difference display shows the results of this calculation:

If the results oscillate around the zero line, no differences can be found there. The “peaks” indicate at which wavelengths the samples differ how much.

What is the difference visualization used for?

The difference display shows even the smallest differences in the spectra, which could not have been detected only by the raw spectra. This representation is used for:

  1. Detecting whether differences exist in the raw spectra
  2. Recognize at which wavelengths differences exist. This gives information about different additives etc.

In the area of batch control, for example, the differences in two batches can be identified quickly and easily. First indications “where” the different samples differ can be found. This is the first step to explain changed behavior in a new batch due to differences in chemistry and to take appropriate countermeasures.

The difference display shows even the smallest differences at a single point in the NIR spectrum.

The difference display shows significant differences in the spectra in the range of 1,100 nm to 1,200 nm. The sample in green further differs from all other samples almost over the entire spectrum.

The example of Huaxia and Richon shows clear differences in the range 1,120 nm and 1,150 nm.

How does the cluster view work?

The cluster display, on the other hand, shows whether different samples differ from each other in sum. Small differences in the NIR spectrum at certain wavelengths are not taken into account.

For the cluster display, the raw spectra are searched for similarities using a mathematical transformation. If the raw spectra are similar, this transformation produces similar results. Similar raw spectra are displayed as “points” in the cluster display, which are spatially close to each other.

What is cluster analysis used for?

Cluster analysis is mainly used for the analysis of large amounts of data, i.e. many raw spectra. Here, commonalities can be found that have not been shown in the representation of the raw spectra, nor in the difference representation.

Furthermore, the cluster view is a very simple data visualization which is also well suited to enable non-experts to identify new samples.

The cluster representation confirms that the sample in green is substantially different from all others. The sum of the differences in the samples purple and blue leads to a clear separation in this representation. The samples in yellow and red are identical material.

Automatic identification with reference data


The automatic comparison includes the entire spectrum and thus also detects deviations in the boundary areas or with secondary substances. In the case of a match, the degree of agreement is precisely indicated.

After you have imported reference plastics or measured them yourself, the program is ready to recognize the scanned plastic.

Fast plastics detection via comparison measurement

Data of the new measurement are first processed and then compared with the data sets available in the reference database by means of statistics. Via a message window you can find out how high the correlation with existing plastics is.

You can add the identified plastic to the reference database by specifying the chemical name, thus improving the data basis for future measurements. It is also possible to add the trade name for easier readability. To record all information about the material, you can enter the batch, the supplier and a further description.

In addition, the individual measurements can be combined into groups. So “Manufacturer A – Measurement 1” and “Manufacturer A – Measurement 2” simply become “Manufacturer A”. This is made possible by the open JDX data format.

This is especially useful if batch control has shown that the new delivery is within the given specifications. The current measurement can then also serve as reference for future measurements.

The measurement result can now be recorded in a report, together with the batch number and other information, assuming you use the Reporting module.