Meaning
Mathematical algorithm smooths a set of data points by weighting the values according to a normal distribution curve. Applying a gaussian low-pass filter to the surface profile of a moulded part removes high-frequency noise and short-wavelength irregularities that might obscure the underlying shape. This technique is standard in metrology for calculating roughness parameters without the interference of measurement artefacts.
Signal Smoothing
Filtering works by convolving the raw surface data with a bell-shaped weighting function. A gaussian low-pass filter suppresses the sharp peaks and valleys caused by tool marks or electrical noise in the sensor. The result is a smoother line that represents the general waviness and form of the part surface.
This modified profile allows for more consistent comparisons between different batches of parts produced from the same cavity.
Cutoff Selection
Width of the gaussian curve, known as the cutoff wavelength, determines which features are kept and which are removed. In a gaussian low-pass filter, a larger cutoff removes more detail, while a smaller cutoff preserves the fine texture of the polymer surface. Choosing the correct cutoff is essential for ensuring that the measurement reflects the functional requirements of the part, such as seal integrity or aesthetic appearance.
Profile Accuracy
Filtered data allows for the calculation of the ra or rz values used in quality specifications.
