Abstract
Gaussian Filter (GF) to extract the roughness profile from a primary profile suffers strongly from the outliers mingled in the data. To solve this problem, though several researches to put this method robust against noise have been proposed, the calculation cost becomes large due to the large number of the repetitive operations. We proposed in the past a Robust GF by using Fast M-estimation method to reduce the fatal computing cost, and we noted its good property that it could provide completely the same result of the original GF method using no outliers in the data. We propose a new method that extended this Robust GF from one dimensional to two dimensional expression for the further applications of 3D roughness profile and confirm the performance of the method.
