Measurement noise

Measurement noise refers to random fluctuations added to the true signal, making it harder to accurately determine the measured value. It is present in all measurement systems and affects the consistency and reliability of the results.
Noise appears as small variations in the measured data that are not related to the actual surface or object being measured. It limits repeatability and contributes to measurement uncertainty.

Measurement noise can originate from different sources, including:

  Instrumental noise: electronic noise, detector limitations, or imperfections in the measurement system.

  Environmental noise: temperature changes, vibrations, acoustic interference, or electromagnetic fields.

Noise affects the quality and interpretation of measurements in several ways:

  Reduces the signal-to-noise ratio (SNR), making small features harder to detect.

  Decreases repeatability and precision.

  Limits the ability to measure very fine surface details.

  Can affect the accuracy of calculated surface parameters.

Noise is typically evaluated by measuring a stable reference surface multiple times and calculating the variation (often as standard deviation or RMS). It can also be determined by performing two measurements of a mirror (or a similar smooth reference surface) and subtracting them to obtain Sq, which reflects the system’s intrinsic noise.

To reduce noise, several strategies can be used:

  Averaging: combining multiple measurements to reduce random fluctuations. This can be done either by averaging repeated full measurements or by averaging multiple images within a single scan.

  Filtering: applying filters (S-filters, L-filters, F-operators in ISO 25178) to separate noise from meaningful data.

  Isolation: minimizing vibrations, temperature changes, and external interference.

  Optimized instrumentation: using high-quality sensors and proper measurement settings.