A New Data Reduction Approach in Coordinate Metrology Using Perona-Malik Anisotropic Heat Diffusion

Abstract

The uncertainty of Coordinate metrology systems is severely affected by the number and location of the sampled discrete points. Also nowadays with the trend for using optical methods which provide a vast number of measured points, data reduction approaches can be used to select the most effective measured points from the cloud of data points to evaluate the geometric and profile deviations. The adopted sampling method affects the metrological analysis and change the computational costs within this type of applications. In this paper, the geometric deviation zone is evaluated by using the Perona‐Malik anisotropic heat diffusion to decrease the number of unimportant measured points. The Euclidian distances of the measured points from the ideal surface is considered as the independent variable in the diffusion equation. By discretization, the location term is discretized by finite difference and the temporal term is estimated by forward Euler method. The results represents the capability of the proposed methodology to reduce the computational cost significantly without imperilling the accuracy of the deviation zone estimation.