Efficiency of Adaptive Sampling in Surface Texture Measurement for Structured Surfaces

Adaptive sampling has been used as an efficient sampling strategy in metrology for many years. However, most of the research into adaptive sampling has concentrated on the area of form measurement by coordinate measuring machines. This paper will discuss the use of adaptive sampling, primarily to reduce the measurement time when making high density areal surface measurements, for instruments that measure surface texture using scanning mechanisms, such as stylus profilometers. Specifically, the paper will concentrate on a common adaptive sampling method known as indirect sampling and a modified version is proposed in this paper. Simulated sampling analyses and three typical micro-scale structured surfaces (a vee-groove-like surface, a rectangularly tessellated micro-lens array and a MEMS device surface) are used as test cases. Using the tensor product order 2 B-spline reconstruction, the root mean square deviation from the original surface is calculated. By comparison with the widely used uniform sampling technique, the performance analysis results show that the modified adaptive sampling has advantages by improving the measurement accuracy, reducing the data size and reducing the measurement duration.