摘要
针对依赖人工进行复杂金相组织分析智能化程度低、精度不高的状况,提出了一种基于改进标记的分水岭算法的金相组织智能检测方法。该方法首先引入多尺度Retinex等算法对金相图像预处理,有效提升金相图像质量;其次,依据拉普拉斯锐化等获得晶界基本轮廓;最后通过改进的自适应标记的分水岭算法,有效地解决了金相组织边界模糊、中断等问题,实现了精确且完整的金相组织检测。对比实验表明:在边缘检测准确度、晶粒数检测以及检测时间等方面都表现优异。该方法不仅可以解决人工分析受主观判定影响大、耗时耗力的问题,也提升了现有方法对复杂的金相组织检测精度低的问题,能够满足晶粒度等金相组织特征参数智能检测需求。
Aiming at the current problems that the complex metallographic structure analysis mainly relies on manual work,the degree of intelligence is low,and the precision is not high,this paper proposes an intelligent metallographic structure detection method based on the marker-improved watershed algorithm.Firstly,by introducing multi-scale Retinex and other algorithms,the method preprocesses metallographic images,which improves the quality of metallographic images effectively.Secondly,the basic outline of grain boundaries can be obtained through Laplace sharpening.At last,by self-adaptive marker-improved watershed algorithm,the method effectively resolves the problems of blurred and interrupted boundary of metallographic structure and realizes accurate and complete metallographic structure detection.The comparative experimental results show that it is superior to the existing methods in terms of edge detection accuracy,grain number detection,and detection time.This method can not only solve the problem that manual analysis is subject to great subjective influence and is time-consuming and labor-intensive,bust also improves the problem of low detection accuracy of complex metallographic structures of existing methods,and can meet the needs of intelligent detection of metallographic structure characteristic parameters such as grain size.
作者
张利欣
孙涵
尧昊天
边胜琴
ZHANG Lixin;SUN Han;YAO Haotian;BIAN Shengqin(School of Automation and Electrical Engineering,University of Science and Technology Beijing,Beijing 100083,China;School of Computer and Communication Engineering,University of Science and Technology Beijing,Beijing 100083,China)
出处
《实验室研究与探索》
CAS
北大核心
2022年第8期1-4,101,共5页
Research and Exploration In Laboratory
基金
国家重点研发计划项目(1796JH0115)
北京科技大学本科教改项目(JG2019M24)。
关键词
自适应标记
金相组织
图像分割
智能检测
adaptive mark
metallographic structure
image segmentation
intelligent detection
