活性碳纤维材料中不规则孔结构的图像定量分析方法研究

项目名称： 活性碳纤维材料中不规则孔结构的图像定量分析方法研究

项目编号： No.50803016

项目类型： 青年科学基金项目

立项/批准年度： 2009

项目学科： 机械、仪表工业

项目作者： 朱煜

作者单位： 华东理工大学

项目金额： 20万元

中文摘要： 本项目利用计算机数字图像处理与分析手段对多孔(孔径>50nm和孔径<2nm)的活性碳纤维材料(ACF)表面微观结构进行了细致有效的分析。SEM等设备获取的显微材料图像信噪比低，微观表面孔结构复杂。本项目结合显微图像处理技术对ACF不规则孔结构进行了大量的定量分析研究，取得了一些成果，主要包括：（1）基于聚类分析方法，提出了材料表面孔洞的尺寸、分布和特定规律性特征的识别和表征方法；（2）研究了SEM显微图像的有效目标分割问题；（3）提出了基于模式识别理论的人工鱼群方法进行目标孔的分割处理；（4）将基于统计计算的隐马尔科夫模型算法应用于微观材料图像目标识别研究；（5）提出了基于孔的方向切线信息的目标分割与计算方法，能够精确获得孔的各类统计参数特征，为ACF的可控制备提供理论依据。本项目结合ACF材料图像，建立了算法软件仿真定量分析系统，对各类算法的实现及对比计算提供了分析平台，成功地为材料实际制备和应用奠定了数据分析的基础。同时，在信息处理技术与材料表面分析这一交叉研究领域进行了有意义的探索。

中文关键词： 活性碳纤维；图像处理；模式识别;孔

英文摘要： In this project,digital image processing methods were successfully used for the surface morphology analyzing of activated carbon fabric (ACF) material with mesopores (2 nm &lt; pore size &lt; 50 nm) and macropores (pore size &gt; 50 nm). The images captured from SEM or TEM are with very low SNR and complx porosity structure. Image processing methods disccused in this project can provide effective and quantitative analysis for the irregular pore structures of ACF material. We obtained some interesting and novelty results. The main works include:(1)based on cluster algorithms, the size, distribution and regularity features of the pores were characterized and analyzed quantitatively.(2) suitable image segmentation methods was proposed for SEM micro-image preprocessing.(3) some effective image segmentation methods were proposed,such as Artificial Fish Swarm algorithm,Snake and etc.(4) Hidden Markov Random Fields model was used for micro-image proccessing and recognition.(5)a tangential direction edge detection method was first proposed for pore's parameter extrection and quatitatively analysis. All the calculated results provide the theoretical basis for controllable manufacturing of porosity ACFs. The simulation software platform was establised to verify various algorithms. All the results will play important role in both theoritical and real applications of the novel ACF materials.This work is a meaningful exploration in the cross-cutting areas of information technology and material surface analysis.

英文关键词： activated carbon fabric; image processing; pattern recognition; porosity