National Physical Laboratory

NPL Reference Stage Graticules

Reference Stage Graticules
Image courtesy of iStockphoto

Reference stage graticules patterns on standard
Patterns on the standard

This calibration artefact contains four test areas comprising a 400 µm square grid, a 20 x 17 monosize array of 15 µm diameter spots, a Root-2 array of spots from 3 µm to 48 µm diameter, and a log-normally distributed array of 100 spots ranging from 4.5 µm to 27 µm diameter. It is ideally suited for calibrating image analyser systems and can also be used as a high precision stage micrometer.

NPL Reference Stage Graticule

The Reference Stage Graticule (RSG) is ideally suited for calibrating image analyser systems and can be used as a high precision stage micrometer.

It contains four test areas comprising:

  1. A 400 µm square grid
  2. A 20 x 17 monosize array of 15 µm diameter spots
  3. A Root-2 array of spots from 3 µm to 48 µm diameter
  4. A log-normally distributed array of 100 spots from 4.5 µm to 27 µm diameter

The reference stage graticule is a 75 mm x 25 mm chrome-on-glass slide and made to NPL specifications.

The centre of the graticule has a clear window approximately 19 mm × 19 mm. The test areas described below are positioned centrally within a clear window and with a cross hair at the centre.

Details of the standard

The grid

The grid shown in Figure 1 is used to calibrate the image analyser and to detect any gross image distortion. An appropriately sized square should be imaged on the screen and the analyser calibrated across the square in the normal way. The squareness of the calibration can be checked by using the other two sides of the square (on some analysers calibration in a second direction is not possible and a variable frame can be used). Many image analysers can produce a software generated grid. If this grid is superimposed on the image of the graticule grid, any large-scale image distortions will become evident.

Grid (400 µm square)Figure 1: Grid (400 µm square)

The monosize array

The monosize array shown in Figure 2 is used to check for localised distortions of the image. These distortions are quite common at the edge of the field of view and a knowledge of where and by how much the scaling breaks down allows determination of the usable measuring area of the image. The slide is positioned so that the 20 x 17 array of spots fill the screen. The height and width of the spots are then measured and either by printing out these sizes with the spot position or by labelling the image on the screen if this is possible, the deviation in the sizes of the spots at the edges can be seen by comparison with spots in the middle of the screen. Measurements must be made rather than direct visual observations because screen distortions may affect judgement.

Monosize Array (15 µm spots)Figure 2: Monosize Array (15 µm spots)

The root-2 array

The root-2 array shown in Figure 3 is used to determine the threshold level required to measure the spots correctly. The image analyser works by the user choosing a grey level at which anything darker is spot and anything lighter is background. As the edge of a spot is not a clean edge but a blurred one due to the limited resolution of the camera and the physical limits of optical microscopy, the specified threshold level affects the measured size of the spots. This array of spots provides a useful research tool for investigating the effects of background lighting, detect level, focus etc. on measured spot size.

Root-2 Progression (3 µm to 48 µm)
Figure 3: Root-2 Progression (3 µm to 48 µm)

The log-normal array

The final test area is the log-normally distributed array shown in Figure 4. This is an idealised distribution of maximum dynamic range for a full screen and is used as a final check on the analyser when all other variables have been corrected or evaluated. The 100 spots should be arranged to fill the screen prior to measurement. Using the software that may be provided with the image analyser, the mean and standard deviation of the log-normal distribution can be determined and compared with the certified values.

Log-Normal Distribution (4.5 to 27 µm)Figure 4: Log-Normal Distribution (4.5 to 27 µm)


The measurement uncertainty can be as low as ±0.1 µm. All uncertainties have been calculated with 95% confidence limit and it is a UKAS-accredited service.


For further information or to discuss your particular calibration requirements, please contact:

Customer Service tel: +44 20 8943 6245

Last Updated: 3 Jan 2019
Created: 1 Feb 2011


Please note that the information will not be divulged to third parties, or used without your permission