Overview
ISO 16610-31:2025 - "Geometrical product specifications (GPS) - Filtration - Part 31: Robust profile filters: Gaussian regression filters" - specifies a robust, nonlinear filtration method for surface profiles. The standard defines robust Gaussian regression filters that separate large‑scale and small‑scale lateral components of surface profiles, reducing the influence of protruding dales and hills. It covers both open and closed profiles (including roundness filtering) and provides continuous filter definitions and a normative iterative solution.
Key topics
- Robust Gaussian regression filter: regression filter based on a Gaussian weighting function combined with the biweight function of Beaton and Tukey (M‑estimation) to limit the influence of outliers such as protruding hills and dales.
- Separation of components: explicit definition of how to obtain large‑scale (form/ waviness) and small‑scale (roughness) lateral components while allowing reconstruction of the original profile.
- Open vs closed profiles: treatment for finite open profiles and closed (periodic) profiles; concepts extendable to generalized closed profiles (e.g., re‑entrant features).
- Nesting index and regression degree: use of a nesting index N (instead of a linear cut‑off wavelength) as the primary filter parameter and specification of the regression degree p for local polynomial modelling.
- Nonlinearity and transmission: because the filter is nonlinear, traditional Fourier‑based transmission characteristics are not directly applicable; the document discusses transmission behavior and practical selection of nesting index values.
- Implementation details: filter equations, iterative solution for continuous filters, filter designation rules, and informative annexes with examples and matrix‑model relationships.
Applications
- Surface metrology and profilometry: preprocessing of measured profiles before roughness and waviness parameter calculation.
- Roundness and form analysis: robust roundness filtering on closed profiles where protrusions or re‑entrant features exist.
- Manufacturing quality control and inspection: improving stability of surface‑parameter extraction when measurements contain scratches, steps or extreme features.
- Software and instrument development: reference for implementing robust filtration algorithms in profilometers, surface analysis tools, and metrology software.
Who should use this standard
- Metrologists and quality engineers working with surface texture and form measurement.
- Instrument and software vendors implementing profile filtering algorithms.
- Laboratories performing surface characterization and compliance testing.
- Standards committees and technical specialists aligning GPS filtration practices.
Related standards
- ISO 16610 series (Parts 1, 20, 21, 22, 30) - overview, linear filters (Gaussian, spline), and robust filter basics.
- ISO/IEC Guide 99 (VIM) for metrology terminology.
Keywords: ISO 16610-31:2025, robust Gaussian regression filter, filtration, GPS, surface profile, nesting index, regression degree, biweight function, profilometry, roundness filtering, surface metrology.