纺织品图像增强能够突出其纹理特性,便于纺织品的人工检测和机器视觉检测。提出一种在非局部均值滤波(NLM)框架下的纺织品图像纹理增强方法。纺织品图像具有规则周期的纹理,存在大量的冗余信息,NLM可利用这一特性来增强图像的纹理信息。但由于纺织品图像结构复杂且存在噪声,导致在NLM中相似性的度量不够准确。为解决这一问题,通过采用主分量分析(PCA)将纺织品图像分解为图像信息分量和噪声分量,并去除各分量间的相关性,来提高纺织品纹理间相似性度量的准确性。实验结果表明,本文方法比现有的纺织品图像纹理增强方法的增强效果有显著提高。

Textile image enhancement aims to extract textural features of textiles which facilitates manual testing and machine vision inspection of textile.In this paper,a method for textile image enhancement is proposed in the framework of non-local means (NLM) filtering.Due to the periodic nature of textiles,there exists a lot of redundant information in textile images which can be used to enhance the texture information.However,the complex structures of textile images as well as the presence of image noise tend to distort similarity measures of the NLM algorithm.To solve this problem,principal component analysis (PCA) is used to decompose textile images into information components and noise components,and remove the correlation between the components for improving the accuracy of the similarity measure between textile textures.The experimental results demonstrate that the proposed method has substantially improved the performance of texture enhancement relative to the existing texture enhancement methods.