Abstract

This paper presents a hierarchical-information-based characterization method for multiscale structured surfaces. The method makes use of a priori information of the multiscale surface and a hierarchical segmentation-registration (HSR) algorithm to firstly divide the measurement dataset into segments, which are then registered with the designed geometry of the microstructure. The registered information is used to reconstruct the large scale topography. Multiple scales of surface information are decomposed hierarchically by the HSR algorithm. Hence, the errors in different scales are determined. Simulation and experimental results show that the hierarchical-information-based characterization method is accurate and effective in the characterization of multiscale structured surfaces.