黄乾玮, 张团善, 周玲, 汤锋, 李乐乐.基于彩色图像处理与EDLines的数码印花缺陷检测系统[J].轻工机械,2021,39(4):74-79 |
基于彩色图像处理与EDLines的数码印花缺陷检测系统 |
Digital Printing Defect Detection System Based on Color Image Processing and EDLines |
|
DOI:10.3969/j.issn.1005 2895.2021.04.014 |
中文关键词: 缺陷检测 颜色空间 颜色补偿 EDLines |
英文关键词:defect detection color space color compensation EDLines |
基金项目:国家自然科学基金(61701384);西安市现代智能纺织装备重点实验室资助项目(2019220614SYS021CG043)。 |
|
摘要点击次数: 518 |
全文下载次数: 715 |
中文摘要: |
针对数码印花生产过程中由于喷头堵塞、电机偏差引起的PASS道缺陷问题,课题组设计了一套基于彩色图像处理与EDLines的数码印花缺陷检测系统。首先构建颜色补偿矩阵覆盖无关背景花案,增强缺陷与主色间差异性;然后分别在HSI颜色空间3通道采用自定义线型滤波锐化感兴趣区域并基于EDLines实现缺陷配准,并将3通道缺陷匹配结果进行区域融合和形态学处理;最后根据水平投影实现PASS道缺陷定位。实验结果表明:检测系统对印花织物表面缺陷的检测准确率达到96%以上,满足实际检测要求,为数码印花缺陷检测提供新的方法。 |
英文摘要: |
A digital printing defect detection system based on color image processing and EDLines was designed to solve the problem of PASS defect caused by nozzle blockage and motor deviation in digital printing production. Firstly the color compensation matrix was built to overlay the irrelevant background pattern and enhance the differences between defects and the main color. Then the region of interest was sharpened in HSI color space three channel by using custom linear filtering and the defect registration was realized based on EDLines. The three channel defect registration results were regional integrated and morphologically processed. Finally, PASS channel defect location was realized according to the horizontal projection. The experimental results show that the detection accuracy of printed fabric surface defects can reach above 96%, which meets the actual detection requirements and provides a new method for the detection of digital printing defects. |
查看全文 查看/发表评论 下载PDF阅读器 |
关闭 |