应用轮廓线拟合提取蓝印花布图案基元  

Extraction of image elements for blue calico based on contour fitting

作　　者：贾小军[1] 邓洪涛[2] 滕姿[1] 曾丹[3] JIA Xiaojun;DENG Hongtao;TENG Zi;ZENG Dan(College of Mathematics,Physics and Information Engineering,Jiaxing University,Jiaxing,Zhejiang 314001,China;College of Design,Jiaxing University,Jiaxing,Zhejiang 314001,China;Key Laboratory of Specialty Fiber Optics and Optical Access Networks,Shanghai University,Shanghai 200072,China)

机构地区：[1]嘉兴学院数理与信息工程学院,浙江嘉兴314001 [2]嘉兴学院设计学院,浙江嘉兴314001 [3]上海大学特种光纤与光接入网重点实验室,上海200072

出　　处：《纺织学报》2018年第8期150-157,共8页Journal of Textile Research

基　　金：国家自然科学基金项目(61572307)

摘　　要：针对蓝印花布的数字化传承与创新问题,提出一种基于轮廓线拟合的蓝印花布图案基元的提取方法,作为蓝印花布设计创新的素材基础。通过对获取的蓝印花布数字图像进行预处理,包括灰度化、去噪和二值化,得到能够反映图案基元特征的二值图像;采用轮廓跟踪法获取图案基元的初始轮廓,并对轮廓进行拟合优化获取理想的基元轮廓;最后提取出蓝印花布的图案基元,并生成可独立读取的图案基元子图。结果表明,利用基于面积及轮廓线像素的评价指标,提取的图案基元准确,轮廓光滑、完整,相对于其他几种轮廓提取方法,可获得更有效的图案基元轮廓以及执行效率。In view of the problem of digital inheritance and innovation for blue calico,a method to extract image elements for blue calico based on contour fitting was presented. These image elements were the material foundation of blue calico design and innovation. Firstly,the acquired blue calico digital images were preprocessed, including graying, denoising and binarization, to obtain binary images reflecting the features of the image elements. Secondly,the initial contours of image elements were obtained by using contour tracking method,and the contours were fitted and optimized to get the ideal contours of image elements. Finally,the image elements of the blue calico were extracted and the sub images of image elements were generated,and they could be accessed independently. The experimental results show that using the evaluation indexes based on the area and contour pixels,the extracted image elements are accurate,and the contours are smooth and complete. More effective contours of image elements and execution efficiency are obtained than the other contour extraction methods.

关 键 词：蓝印花布 图案基元 轮廓提取 评价指标 

分 类 号：TP391.7[自动化与计算机技术—计算机应用技术]