摘要
提出一种新的无标记印刷图像质量检测方案,采用HALCON编程实现,主要包括配准区域选取、图像配准和缺陷检测等关键步骤.首先根据印刷品图像的区域特征自动选择配准区域,然后使用基于形状的图形配准算法实现快速的图像配准,最后利用改进的差影匹配算法完成缺陷检测.实验结果证明,该方法能从无标记印刷品图像中自动选取配准区域,并且快速、精确地完成配准,有效地检测出缺陷.
A new approach of quality detection of the printing image based on HALCON is proposed. Our design is based on HALCON,a machine vision software,which includes key processing steps such as chosing matching regions,shape-based matching,defects detection and so on. First,the matching region based on the region features is chosen,such as area,rectangle descriptor and circular descriptor.Then,fast image matching in the approach of the shape-based matching is realized. At last,the improved difference-matching algorithm to detect the printing defects is used. The result of the experiment shows that our approach can choose the proper matching regions from the non-mark printing image automatically,match the images with high accuracy and speed,and detect the printing defects efficiently.
出处
《汕头大学学报（自然科学版）》
2011年第2期63-68,80,共7页
Journal of Shantou University：Natural Science Edition
基金
轻工装备产学研示范基地开放课题资助项目
