Correlation Image Analysis of Surface Roughness

Autor: Abul’hanov, S.R., Skuratov, Dmitrii Leonidovich, Khaimovich, Alexander I.
Zdroj: Key Engineering Materials; July 2017, Vol. 746 Issue: 1 p296-304, 9p
Abstrakt: The authors proposed a procedure for determining parallel matchmarks in the halftone image of the rough surface. The analysis of micron-scale asperities was carried out in black and white. Relationships between coordinates of variable pixel-value sections of the microasperities under investigation were identified using the correlation analysis. The authors determined the asymmetrical distribution of various grayscales in the image of the polished surface. For that reason, the Spearman’s rank correlation coefficient was found to identify microrelief regularities that enabled to determine peaks of parallel matchmarks available on the rough surface.It is necessary to apply special filtration in order to determine periodicity of linear matchmarks which will remove point inclusions from the surface image (carbides, impurities, impregnated particles, etc.). Small inclusions in the image of the surface being analyzed are considered to be the “noise” while conducting the correlation analysis.The findings are relevant for quality control of optical surfaces. The developed program for calculating the correlation coefficient may be built into a product automatic control system used in bulk production.
Databáze: Supplemental Index