基于阈值分割及分形特征的纸病图像识别算法研究

Identification Algorithm of Paper Defects Based on Threshold Segmentation and Fractal Characteristics

中文关键词:  灰度特征  阈值分割  分形盒维数  纸病识别

Key Words:gray characteristics  threshold segmentation  fractal box dimension  paper defects recognition

中文摘要:

      针对脏点、孔洞、褶子和裂口等常见的典型纸病识别问题,在分析纸病图像灰度特征及分形特征基础上,提出了一种基于图像双阈值分割盒维数特征的纸病识别算法,该算法采用灰度阈值分割提取纸病区域及二值数字图像分形盒维数计算结果,确定纸病类型。实验结果表明,该算法识别率较高,且简单迅速。

Abstract:

      Aiming at the identification of typical paper defects such as specks, hole, flaw and goffer, through the analysis of the image gray characteristics and fractal feature, a recognized algorithm of paper defects base on threshold segmentation and fractal characteristics is proposed. The algorithm can determine the types of paper defect, according to image characteristics of paper defect after threshold segmentation, and the box-counting dimension of 2-D digital image. Experiment results show that this approach is efficient, simple and fast for paper defects recognition.