摘要
以一种常见的织物纹理为对象,采用灰度共生矩阵的方法进行纹理分析。介绍了灰度共生矩阵的原理及其特征参数,讨论了纹理的灰度共生矩阵特征参数、像素距离以及图像灰度等级对灰度共生矩阵的影响,确定了区分此类正常织物与带疵点织物纹理的灰度共生矩阵构造方法。针对该类正常织物图像进行纹理分析,特征参数值统计,确定了正常织物纹理像素方向、像素距离以及图像灰度等级。取原始织物图像尺寸为128×128,生成灰度共生矩阵的最佳像素距离为2,经直方图均衡化后,最佳灰度等级为16。实验结果表明,按照该规则生成的6个灰度共生矩阵的特征参数,能够准确的判断此类织物图像是否存在疵点。
AkindofordinaryfabrictextureisstudiedwithGLCMmethod. The theory ofGLCM and its feature parameters are introduced particularly and the influence of building step, building direction and image grey level on parameters of GLCM is discussed in order to make sure a method of GLCM, which can identify faultless fabric texture and the defects. Pixels direction, pixels interval and image gray level of faultless fabric texture are ensured by GLCM. The test indicates that is suitable with the group parameters, such as angular second moment, contrast, correlation, entropy, sum of squares, and inverse difference moment. The better conclusion is 2 ofpixels interval, 16 of image gray level under 128 × 128 fabric image size. The result is the six parameters of GLCM building rules mentioned above describing fabric texture feature can judge defects from fabric image.
出处
《计算机工程与设计》
CSCD
北大核心
2008年第16期4385-4388,共4页
Computer Engineering and Design
