基于复合条件的织物纹理模糊识别

提出一种新的基于复合条件的模糊识别方法完成对织物图像的纹理提取。首先对整个图像灰度级分析,确定行与列灰度波的标准模糊集,然后给出基于灰度大小及像素位置的复合条件隶属度计算公式,对行与列的灰度波进行模糊分类,使其转换为仅有2个灰度极值的纹理灰度波。最后将所有纹理灰度波组合还原成二值纹理织物图像,为后续密度识别、纹理均匀度识别及疵点识别等图像分析提供基础。实验与分析结果显示,本文算法通过对灰度波的模糊划分,提取仅用2个灰度值表示的织物二值纹理图像,形成具有脉冲特征的纹理灰度波,可为后续织物图像各参数分析提供简单有用的织物纹理数据模型。

Fuzzy recognition of fabric texture with double conditions

In order to extract the texture of fabric image,a new fuzzy recognition method on double conditions is proposed.With this algorithm,the gray level of image is analyzed to determine the standard fuzzy set of gray wave in any row and column.And then,the membership function considering different gray values and pixel positions on double condition is given to classify the gray wave in any row and column,which is transformed into the gray wave of texture with only two gray values.And finally,all gray wave combinations of texture are reduced into binary texture image of fabric,thus providing a basis for subsequent image analysis for recognitions of density,texture evenness,and defects.Experiments and analysis show that this algorithm can establish a gray wave with pulse characteristic according to the fuzzy partition of gray wave and binary texture image extraction of fabric with two gray values.Therefore,this algorithm can provide simple and useful data models of fabric texture for further parameters analysis of fabric images.