Quantitative characterization of lamellar and equiaxed alpha phases of (α + β) titanium alloy using a robust approach for touching features splitting

Quantitative characterization of lamellar and equiaxed alpha phases in (α + β) titanium microstructures is needed to investigate how the microstructure affects the mechanical properties. However, the complex touching features between the lamellar and equiaxed alpha of the microstructure make quantitative analysis challenging. To overcome this problem, a four-level procedure for separating touching features, which includes image pre-processing, concave points detection, splitting line determination, and lamellar and equiaxed alpha discrimination, has been proposed to automatically process many microstructures. Moreover, a set of criteria was established for each level to handle various problems caused by the complexity of how lamellar and equiaxed alpha can touch each other. The result on a representative microstructure of a (α + β) titanium alloy, a tri-modal microstructure, shows that the approach works well and is robust. Based on this effort, the phase distributions of the tri-modal microstructure were quantified through two-point and lineal-path correlation functions based on the newly developed, efficient algorithms. Significant heterogeneity and morphological anisotropy in different directions and ranges of the lamellar and equiaxed alpha were quantitatively measured.