Sampling optimization for 3D surface measurement. Part II: Sampling interval optimization based on cumulative power spectral density analysis

Abstract

Sampling interval is a key sampling parameter in 3D surface measurement, which affects uncertainty and time of measurement. This study aimed to optimize the sampling interval of 3D isotropic and anisotropic surfaces based on cumulative power spectral density (PSD) analysis for the comparability of 3D surface measurement meanwhile ensuring the measurement efficiency. Twenty-four surfaces with expected height distribution and spatial distribution were generated based on Fast Fourier Transform (FFT) and autocorrelation length function. The effects of sampling interval on surface texture parameters and cumulative PSD of isotropic and anisotropic surfaces were investigated under the conditions of redundant, appropriate and distorted sampling intervals, respectively. The results show 3D texture parameters were insensitive to sampling interval, but cumulative PSD significantly depends on sampling interval. Therefore, based on the cumulative PSD analysis, the optimal sampling intervals for isotropic and anisotropic surfaces with different mean surface height ranges were given. It also indicates that the optimal sampling intervals parallel to the texture direction were higher 10–20 times than that perpendicular to texture direction for anisotropic surfaces. In addition, to verify the validity of the recommended sampling intervals, three actual surfaces were machined using different abrasive grit burs and were measured using a non-contact confocal profiler with the optimal sampling intervals recommended. They demonstrate that the recommended optimal sampling intervals could be used as the sampling parameter selection guidance for 3D surface measurement and characterization.