应用方向梯度直方图和低秩分解的织物疵点检测算法

织物疵点检测是织物表面质量控制的关键环节。基于方向梯度直方图(HOG)和低秩分解,提出一种有效的织物疵点检测算法。首先,将织物图像划分为大小相同的图像块,提取每个图像块的HOG特征,并将图像块特征组成特征矩阵,针对特征矩阵构建有效的低秩分解模型,通过方向交替方法(ADM)优化求解,生成低秩阵和稀疏阵;最后采用改进最优阈值分割算法对由稀疏阵生成的显著图进行分割,从而定位出疵点区域。实验结果表明,低秩分解能有效实现织物疵点的快速分离,与已有方法进行对比,本文方法能显著提高复杂织物纹理图像的疵点检测性能。

Fabric defect detection algorithm based on histogram of oriented gradient and low-rank decomposition

Fabric defect detection plays an important role in controlling the quality of fabric surface.An effective fabric detection algorithm based on histogram of oriented gradient (HOG) and low-rank decomposition was proposed.Firstly,the test fabric image was divided into image blocks with the same size.A feature matrix was generated by extracting the HOG feature of each block.Secondly,an efficient low-rank decomposition model was constructed,and alternating direction method (ADM) was adopted to decompose the feature matrix into a low-rank matrix and a sparse matrix.Finally,the saliency map generated by sparse matrix was segmented via an improved optimal threshold algorithm to localize the defect.The experimental results show that the proposed method can sufficiently improve the defect detection performance of complicated textile texture patterns.