摘要
针对当前纸病在线检测系统实时性强和信息量大的特点,提出了一种高效、灵活的基于几何及灰度特征的纸病检测算法。首先采用邻域均值法对纸病图像去噪,然后根据灰度直方图选取合适阈值将图像二值化,并运用边界跟踪法检测出纸病边缘,最后提取出纸病的几何及灰度特征并分析其特征量将纸病分类。按照该算法依次对常见5种纸病进行检测,结果表明,基于几何及灰度特征的纸病检测算法能够准确地检测并分类常见纸病。
Considering the characteristics of strong real-time and a great deal information of current paper defect on-line detecting system,an efficient and flexible algorithm of paper defect detection based on geometric and gray feature is proposed in this paper.The paper defect image is de-noised by neighborhood averaging method at first.And then binary image is obtained through selecting appropriate threshold according to gray histogram.At last paper defect edge is detected by boundary tracking method.Geometric and gray features of paper defects are extracted and analyzed so as to classify them.Experiment was carried out to verify this algorithm for five common paper defects.The result showed that common paper defects can be accurately detected and classified through the algorithm of paper defect detection based on geometric and gray feature.
出处
《中国造纸》
CAS
北大核心
2011年第9期50-53,共4页
China Pulp & Paper
基金
陕西省教育厅科研专项基金(2010JK420)
陕西科技大学校博士科研启动基金(BJ10-05)
陕西科技大学校级学术骨干培养计划(2010)资助
关键词
纸病
在线检测
几何特征
灰度特征
特征提取
paper defect
on-line detection
geometric feature
gray feature
feature extraction
