Example 2: Papers

This example shows that different paper types can be distinguished with the NIRScreen. The types of paper used are a handkerchief, a letter, a newspaper and a cardboard box.

Download
Papers.pdf
Adobe Acrobat Dokument 414.9 KB

1. Raw data

  • Calibration measurements are performed as a first step.
  • Each type of paper is measured in three different places with the NIRScreen.
  • This results in 12 calibration measurements in total.
  • The measurement results show that the papers are already clearly distinguishable at first glance.
  • Within the groups, the measurement results differ minimally in height.

Raw data:

Handkerchief (red), Letter (blue), Newpaper (green), Carton (orange)


2. Processed data

  • As a second step, the raw data are processed with an algorithm to correct scattering effects within the groups.
  • After processing the data, the measurements within a group are barely distinguishable.
  • However the measurements between the different types of paper are slightly different.
  • It can therefore be assumed that the paper types can be differentiated.

Processed data:

Handkerchief (red), Letter (blue), Newpaper (green), Carton (orange)


3. Principal Component Analysis

  • A principal component analysis is performed as a third step to make sure that the paper types are truly distinguishable.
  • Looking at the results of the principal component analysis, it is noticeable that the measurements can be divided into four different groups.
  • This confirms the impression of the processed data that the paper types can be distinguished.

Result of Principal Component Analysis:

Handkerchief (red), Letter (blue), Newpaper (green), Carton (orange)


4. Statistical model

  • The final step is to generate a statistical model that will allow you to predict the paper type.
  • These statistics models are stored in the cloud and can be downloaded from a NIRScreen.
  • Based on model parameters, it can be read that the paper grades can be distinguished well.
  • The statistical model also shows how strong the influence of wavelengths is on the differentiation of papers (see graph).

Influence of the wavelengths on the differation of paper types