基于数字图像处理的浆纱截面参数分析

【作者】 姜珊；

【导师】 王跃存；

【作者基本信息】 天津工业大学 ， 信号与信息处理， 2008， 硕士

【摘要】 传统的浆纱截面图像分析主要依靠人工检测分析,效率较低。将数字图像处理技术应用于浆纱切片图像的自动分析,改进了纺织行业中传统的手工操作流程,提高了工作效率,对实现纺织产品检测的自动化、智能化具有现实意义。目前常用的浆纱截面图像分析方法主要是基于投影仪的人工分析和基于图像分析软件的半自动分析。课题设计中引入数学形态学运算,考虑通过不同型号的视频显微镜采集的不同种类浆纱截面图像的特点,提出了一种基于数学形态学的浆纱切片自动分割算法,包括浆纱截面的分割,浆纱被覆与浸透部分的分割等。该方法在对浆纱切片图像进行预处理后,利用灰度填充算法获得浆纱截面在图像中的位置,并用闭合、面积滤波、孔洞填充等一系列处理分割出完整的浆纱截面以及浆纱截面的被覆与浸透部分。之后提出了一种基于特征点连接的算法来确定纱线的边缘,实现了浆纱截面被覆与浸透部分的分割,计算出了浆膜完整率。最后通过对浆纱截面总面积以及被覆与浸透部分面积的计算得到了浆纱的被覆率和浸透率。在设计中充分考虑了浆纱截面图像的特点,严格按照人工分割浆纱截面被覆与浸透部分的标准来分割图像,获得了比较理想的处理结果。本文对多幅浆纱截面图像进行了实验,结果表明算法可以有效地进行浆纱截面参数的分析,处理速度快且准确率较高。更多还原

【Abstract】 In the textile industry, the analysis of sizing section images is usually done by workers, so the analysis procedure is very time consuming and inefficiency. In order to improve work efficiency, Digital Image Processing is utilized to achieve automatic analysis of sizing section images. It plays an important role to improve automatization and artificial intelligence in textile production. Nowadays the realization of analysising sizing section images commonly based on projection equipments and image analysis software.This thesis proposes a new automatic segmentation algorithm for sizing section images to segment the entire sizing section, the coating part and penetration part of them, the mathematics morphology approach was used and the characteristics of different sizing section images got through distinct video microscopes was carefully considered. This approach is based on locating the sizing section in the image which has been preprocessed, and the entire sizing section, coating part and penetration part are segmented after the processing of close, area filtering and filling holes algorithms. Further, an approach based on mark points’ connection is proposed to get the edge of the yarn and separates the coating and penetration part of the sizing section and gets the integrity rate of it. At last, the coating rate and penetration rate are got through computing the areas of entire sizing section, coating part and penetration part of the microscopic images. The characteristics of the sizing section microscopic images are carefully considered and the standards of segmentation methods are strictly abided, the results are satisfiable.Experiments are done through many sizing section microscopic images and the results show that the algorithm can be used to process these images in short time and at a high accurate rate.更多还原