构建基于SEM图像的微纳米纤膜三维模型  

Construction of the three-dimensional model of micro-nano membranes based on SEM images

作　　者：叶文丽 王荣武[1] 黄晨[1] 周梦娟 Ye Wenli;Wang Rongwu;Huang Chen;Zhou Mengjuan(College of Textiles,Donghua University,Shanghai 201620,China)

机构地区：[1]东华大学纺织学院,上海201620

出　　处：《产业用纺织品》2020年第12期27-34,共8页Technical Textiles

摘　　要：通过扫描电子显微镜(SEM)获取以非织造材料为基布的静电纺丝纳米过滤膜的SEM图像,采用数字图像处理技术,如中值滤波、Canny算子边缘检测、最佳阈值分割等算法,对该微纳米纤膜SEM图像进行预处理,提取SEM图像中的纤维边缘信号,在此基础上,通过霍夫变换将纤维边缘像素信息直线化,并采用配对及偏移算法,获得每根纤维的中轴线,再利用纤维二维像素、坐标信息及层叠关系,将在平面上直线化的纤维中轴线三维化,使每根纤维的中轴线都有其体积深度信息,由此构建基于SEM图像的纤维三维层叠结构简易模型,以探索微纳米纤膜的过滤性能。The SEM images of the non-woven fabric-based electrospun nanofiltration membranes were obtained through a scanning electron microscope,micro-nano membrane images were preprocessed with digital image processing techniques such as median filtering,Canny edge detection,threshold segmentationetc to extract the fiber edge signals.The fiber edge pixel information was linearized by Hough transform on this basis,and then the pairing and offset algorithms were used to obtain the unique central axis of each fiber.Using the two-dimensional pixels,coordinate information and the lamination relationship of the fiber basis of extracting edge signals of fibers in the SEM image,the central axis of the fiber linearized in the plane was made three-dimensional,that was,each fiber axis had its volume depth information,and thus a simple model of the three-dimensional layer structure of the fiber was obtained in the SEM image,to explore the filtration performance of micro-nano fiber membranes.

关 键 词：数字图像处理技术 微纳米纤膜 边缘信号 霍夫变换 三维层叠结构简易模型 

分 类 号：TS171[轻工技术与工程—纺织材料与纺织品设计]