基于机器视觉的圆网印花机对花检测系统  

Register detection system of rotary screen printing based on machine vision

作　　者：景军锋[1,2] 李光燕[2] 李鹏飞[2] 王刚[2] 

机构地区：[1]西安电子科技大学机电学院,西安710071 [2]西安工程大学电子信息学院,西安710048

出　　处：《电子测量与仪器学报》2011年第12期1066-1071,共6页Journal of Electronic Measurement and Instrumentation

基　　金：陕西省科技厅(编号编:2010K09-17)资助项目;陕西省教育厅(编号:11JK0910)资助项目

摘　　要：针对圆网印花对花检测问题,提出利用机器视觉的方法来检测印染过程中对花误差的大小。首先介绍了印染织物图像的特点,并结合"错花"图像的特点,采用JSEG算法对织物图像进行分割;选取各颜色区域的边缘轮廓作为匹配的特征信息,通过两次基于Fourier-mellin变换的曲线匹配,完成误差的检测。最后给出了"错花"误差的计算结果,表明该方案是可行的,为下一步实际的应用提供了理论依据。Aiming at the register detection problems in rotary screen printing,the machine vision technology is proposed to detect the register error in the dyeing process.Firstly,the characteristics of the printing and dyeing textile image are introduced.Then the JSEG algorithm is used for fabric image segmentation.Combining with the features of ＂ broken pattern＂ image,choosing the edge of the color regions as matching feature information,and through two times of curve matching based on Fourier-mellin transform,the error detection is implemented.Finally,the ＂ broken pattern＂ error is calculated.It shows that the scheme is feasible,and it provides a theoretical basis for the further actual application.

关 键 词：圆网印花机 机器视觉 JSEG FOURIER-MELLIN 

分 类 号：TS194-33[轻工技术与工程—纺织化学与染整工程]