Automatic measurement of grinding-induced white layer in titanium alloys based on image processing technique

摘要

Titanium alloys is a key kind of materials in both aerospace industry and biomedical engineering. However, mechanical grinding process of Ti alloys easily leads to the featureless white layer (WL) extending tens of micrometers beneath the ground surface. Although many relevant studies have been reported so far, the WL quantification in most previous studies was performed by human raw eyes. To fill this gap, this paper proposes an image-processing-based method which can automatically recognise and measure the grinding-induced WL in the Ti alloy workpiece. By comparing the results separately obtained by the proposed method and manual measurement, it shows that the method can recognise and measure the WL region accurately (the maximum absolute error of 0.09?μm) and quickly (1.65?s per micrograph). More importantly, the method does not require any parameter presetting and has good robustness in terms of (i) varied WL region thicknesses, (ii) different workpiece placements and (iii) the interference from the noise pixels. The method therefore is believed to be not only meaningful and helpful to automatically measure the WL thickness in Ti alloys in large quantities, but also transferable to other industrial applications.