摘要
在线彩色印刷品的质量检测系统在采集图像时,由于设备及环境因素的影响,所采集的图像不可避免的混入噪声,并存在旋转、平移和缩放,为消除噪声并使标准图像与待检图像对准,提出一系列去噪、分色以及匹配定位等预处理算法,并进行了理论分析和实验验证,得到了较为理想的处理效果,为下一步的印刷品质量检测奠定了基础。
When images of printing are collected by online quality detection system, noise intermixes into images inevitably, and revolution, translation movement, and zoom exist in images because of equipment and environment. Pretreatment algorithms such as denoising, color separation, matching, and orientation were put forward to eliminate noise and register standard images with collected images. They were analyzed and verified. It was showed that the results are ideal. The purpose of this research is to provide reference for further quality detection of multi-color printing.
出处
《包装工程》
CAS
CSCD
北大核心
2007年第1期64-66,共3页
Packaging Engineering
关键词
匹配
噪声
中值滤波
分色
matching
noise
median filtering
color separation
