基于改进Hough变换的复杂纸页缺陷检测算法

The Complicated Paper Surface Defects Detection Algorithm Based on Improved Hough Transformation

关键词:  Hough变换  纸页缺陷  图像处理  特征提取

Key Words:Hough transform  paper surface defect  image processing  features extraction

摘要:采用传统Hough变换法检测复杂纸页缺陷如褶皱、条痕等直线类特征纸病时，不能有效分辨具有较近斜率或截距的多直线特征。在标准Hough变换的基础上将Hough变换域空间变换为二维图像然后对其进行纸页缺陷特征提取，从而更加精确地检测出褶皱或条痕类纸病。经实验证明，改进的Hough变换在检测精度和时间上都优于标准Hough变换。

Abstract:The complicated paper surface defects such as wrinkle and crease with linear features which have close slope or intercept cannot be distinguished effectively by traditional Hough transform. Based on existing algorithm, the Hough space is transformed into two-dimensional image, and then the paper defects features are extracted, through which the wrinkle and crease surface defects can be detected more accurately. Finally, the experiment proves that the improved algorithm has higher detection precision and shorter time consumption compared with Hough transform.