基于图像处理的包装印刷缺陷检测方法

Package Printing Defect Detection Method Based on Image Processing

目的为了提高包装印刷品缺陷检测的准确度和适用性,基于图像处理设计一种包装印刷缺陷检测方法。方法采用几个关键步骤包括图像配准、配准区域自动选取、缺陷检测等,并根据包装印刷品图像特征选择配准区域,同时给出一种快速图像配准算法,利用改进差影匹配算法实现缺陷检测;基于DSP和FPGA设计控制系统硬件平台,主要包括控制单元、图像采集与处理单元、成像单元等,并进行实验研究。结果所述方法能够准确识别出细微的刀丝、拉条类缺陷。结论该方法具有较高的可靠性、通用性,可实现包装印刷品缺陷的快速检测。

The work aims to improve the accuracy and applicability of package printing defect detection, and design a detection method based on image processing.Several key steps, including image registration, automatic selection of registration areas, and defect detection were adopted.The registration area was selected according to the image features of package printing matter.At the same time, a fast image registration algorithm was presented.Then, the defect detection was achieved on the basis of the improved subtraction matching algorithm.The hardware platform of control system was designed based on DSP and FPGA, mainly including control unit, image acquisition and processing unit, imaging unit and so on.Finally, the experimental study was carried out.Experimental results showed that the described method could accurately identify the subtle defect, such as knife filament and brace.The method has higher reliability and universality, which can realize the rapid detection of package printing defects.