Noise bias removal in profile measurements

In form and roughness measurements, noise may give an offset in a measurement parameter as the noise makes the parameter, e.g. the Ra-value, deviate away from zero. In this paper we propose a method to correct for this noise bias for the roughness parameter Rq which is equivalent to the standard deviation. By considering the decrease in Rq once an average over multiple measurements is made, an unbiased value for Rq is estimated by extrapolation. This principle is extended to obtain a complete 'bias-reduced' profile by considering the change of each Fourier component with averaging. Considering the statistical significance of each Fourier component enables a further reduction. It is shown that using this method for two profile measurements only, the true measurement is approached better than with averaging dozens of measurements. Simulation and measurement examples are shown for roughness and roundness measurements.