基于图像处理技术的多晶体材料金相图像分割算法研究  

Research on Segmentation Algorithm on Metallographic Image of Polycrystalline Material Based on Image Processing Technology

作　　者：王宝珠[1] 李芳影[1] 刘翠响[1] 胡洋[1] 

机构地区：[1]河北工业大学电子信息工程学院,天津300401

出　　处：《铸造》2015年第11期1078-1081,共4页Foundry

摘　　要：选取多晶体材料金相图像为研究对象,针对图像边界不连续、纹路噪声严重等特性,应用数字图像处理技术对多晶体材料金相图像进行处理,实现晶粒的准确分割。首先,对多晶体材料的金相图像进行图像平滑、形态学重建等预处理,然后采用改进的分水岭变换法实现图像中晶粒的分割,为多晶体材料的定量分析和三维重建奠定了坚实的基础。Microstructure segmentation of polycrystalline material, which has the characteristics such as discontinuous boundary and serious grain noise, was realized by capture and processing with digital image technology. First of all, preprocessing the polycrystalline material metallographic image such as image smoothing and morphological reconstruction, then, realizing the image segmentation with the improved watershed transformation. The present work lays a basis for quantitative analysis and 3D microstructure of polycrystal microstructure.

关 键 词：各向异性-中值滤波 形态学重建 脊检测 分水岭分割 

分 类 号：TG115.21[金属学及工艺—物理冶金]