项目名称: 基于信息流、纱疵级数和异纤质数的异性纤维检测算法研究
项目编号: No.51205288
项目类型: 青年科学基金项目
立项/批准年度: 2013
项目学科: 机械工程学科
项目作者: 杜玉红
作者单位: 天津工业大学
项目金额: 25万元
中文摘要: 异性纤维是影响织物质量重要因素。由于异性纤维种类繁多,本项目依据纺纱生产工艺,以纱线形成过程作为异纤信息流走向,研究异性纤维检测相关算法。课题主要研究内容:利用支持向量机研究离散状态下异性纤维静态数学模型;结合纤维信息流时空数据特征,利用随机平稳过程理论建立异性纤维信息流动态模型;采用信息流扁平化理论,分析各类异性纤维信息流内在规律及相关度;研究异性纤维纱疵分级矩阵,建立异纤与纱疵之间关系模型;分析有关异纤纱疵不同信息数据的综合表达与决策分析过程的机制,建立异纤质数的异性纤维特征识别算法。研究基于纱疵的异纤质数,改进小波极值变换,对异性纤维进行识别检测,最后通过实验验证、完善建模方法和模型,完善异纤质数的异性纤维识别算法。本项目旨在研究工艺过程中的异性纤维和纱疵之间的特性规律,为有效控制异纤以及研究检测装置奠定基础。
中文关键词: 异性纤维;图像处理;纱疵;聚类分析;信息融合
英文摘要: Foreign fiber is the main factor affecting the fabric quality. As foreign fiber variety, this project according to the spinning technology, the yarn formation process as a foreign fibers in the flow of information to the opposite fiber detection algorithm, Main research topics : using the support vector machine static mathematical model of foreign fibers in the discrete state; combination of the data characteristics of the fiber flow of information, using a random stationary process theory to establish the dynamic model of foreign fibers in the flow of information; adopting flattening information flow theory, analysis of various types of the foreign fiber and related information flow of foreign fibers; researched foreign fiber faults classification matrix to establish the model of the relationship between the foreign fiber and yarn faults; analyzed the mechanism of expression and decision-making on foreign fiber yarn faults of different information and data analysis process, the establishment of the number of foreign fibers of foreign fiber quality feature recognition algorithm. Basing on the number of foreign fiber yarn faults, improved the extremal wavelet transform, the recognition and detection of foreign fibers, and finally verify the number of experiments to improve the modeling approach and model to impro
英文关键词: Foreign fiber;image processing;Yarn faults;cluster analysis;Intelligence Fusion