A Digital Evaluation Approach for Nodule Size Analysis of Pearlite Mixture Phase Structure in Steel

Abstract

This study employed a digital evaluation approach for determining the nodule size of a no heat-treatment pearlite structure, as well as the grain size of the ferrite phase via an image morphology process methodology. The grain size of steel has great influence on the mechanical properties of manufactured materials, making it a necessary metallographic inspection item for materials. The properties affected by grain size include hardness, failure life, ultimate stress, electric characteristics, and heat expansion rate. In general, grain size inspection has two approaches: manual comparison and digital computation. In manual comparison, the comparison of standard graphics or the observation of an eyepiece inserted on a microscope can be progressed. For digital computation, planimetric and intercept procedures are applied according to international standards. This study thus applied an image processing algorithm to identify the observed image after undergoing seven times of erosion and dilation operations.

Graphite Classification of Gray Cast Iron in Metallographic via a Deep Learning Approach

“…Kiyomura, Wang, and Ogawa et al applied homology and Bayesian optimization of machine learning algorithms to predict the characteristic properties and microstructural optimization of globular cementite in pearlite steel [7]. Our previous research applied image morphology approaches to progress metallographic images of carbon steel and to estimate the grain size of pearlite mixture-phase steel [8]. Tsutsui, Terasaki, and Uto et al used machine learning approaches to utilize feature extraction with texture analysis on several microstructures, including martensite, upper bainite, lower bainite, and other mixture phases, in scanning electronic microscope (SEM) images [9].…”