机器学习在M/A岛和退化珠光体识别中的应用  

Application of machine learning in M/A constituent and degenerated pearlite identification

作　　者：郝英敏 梁晓军[1] 邢钊[1] HAO Yingmin;LIANG Xiaojun;XING Zhao(Research Institute,Baoshan Iron&Steel Co.,Ltd.,Shanghai 201999,China)

机构地区：[1]宝山钢铁股份有限公司中央研究院,上海201999

出　　处：《宝钢技术》2020年第6期34-38,共5页Baosteel Technology

摘　　要：利用机器学习随机森林算法,以高等级厚板钢典型低碳贝氏体组织中M/A岛和退化珠光体为研究对象,探讨随机森林算法在M/A岛识别、进行量化分析的可行性。结果显示,机器学习识别数据基本符合目前对M/A岛和退化珠光体形成规律的认识,随机森林算法可以快速学习并对M/A岛不同特征进行识别,具有较好的效果,为PIDAS中集成组织数据提供了可能性。Machine learning algorithm random forest is used to study M/A constituents and degenerated pearlite in high grade and low carbon bainitic heavy plate steels,and further to explore the feasibility of random forest algorithm in identification in M/A constituents,as well as the data quantitative analysis.The results show that the machine learning results basically conform to the current understanding of the formation mechanisms of M/A constituents and degenerated pearlite,and the random forest algorithm can quickly learn and recognize different features of M/A constituents,which proves to be an effective method,and provides the possibility of integrating microstructure data in PIDAS.

关 键 词：高等级低碳贝氏体钢 M/A岛 随机森林 图像分割 数据分析 

分 类 号：TG142.1+1[一般工业技术—材料科学与工程]