Research On Online Detection Algorithm Of Printed Character Defects Based On Distortion Region Optimization Correction

Printing industry is the main force in the development of China's cultural industry.With the rapid development of the national economy and science and technology,the printing industry is constantly moving toward automation and digitization,and printing products are gradually infiltrating into all aspects of people's lives.In order to improve market competitiveness,printing companies are increasingly strict with the quality of printed products.However,in the printing production process,due to the interference of uncertain factors such as the printing process,equipment and the external environment,the actual printed materials may have character defects such as knife marks,flying ink,drag pens,and broken strokes.These character defects can cause loss of product information,so quality inspection of printed character defects is particularly important.The traditional artificial defect detection method can not meet the modern industrial production demand because of its great limitations and subjectivity.Machine vision technology has gradually replaced manual detection because of its non-contact,high precision and fast speed.It has become the mainstream direction in the field of defect detection.Therefore,the research on the on-line detection technology of machine vision printing character defects proposed in this paper has important practical significance for improving the production efficiency and product quality of printing enterprises.Based on the analysis and summary of the existing character defect detection methods,an online detection system for printed character defects based on distortion region optimization correction is proposed in this paper.According to the specific detection object and testing requirements,the construction of the hardware system for the detection of printed character defects,the research of character defect detection algorithm and the development of software system are completed.The main research work includes the following points:(1)Construction of on-line detection hardware systemThe character defect detection hardware system is divided into four parts: light source,image acquisition module,control module and mechanical transmission module according to the installation structure and function.Through the design and integration of these four parts,a hardware experiment platform based on industrial computer,machine vision imaging system and mechanical transmission device is built.The platform completes the acquisition of printed character images,and uses industrial computer to process the collected character images.(2)Research on printed character defect detection algorithmThe research of character defect detection algorithm includes three parts: preprocessing,defect detection and defect classification.Aiming at the problem of difficult detection of surround characters,a preprocessing method,such as multi-region segmentation,is proposed to detect defects in small-space images,which avoids the problem of low detection accuracy to a certain extent.Aiming at the difficulty that the accuracy of defect detection is limited by the effect of image registration,a printing character defect detection algorithm based on distortion region optimization correction is proposed.By using the bacterial population algorithm to optimize the control points to obtain the best registration effect and optimal difference result.Accurate and efficient defect detection of printed characters is realized.Finally,the SVM support vector machine is used to accurately classify the defects of printed characters.(3)Development of on-line detection software systemThe character defect detection software system is developed on the basis of the MFC class library in the VS2010 development environment using the C++ language.Through the design of image acquisition module,image processing module,interface control and data management module,the detection and classification of printed character defects and the display and storage of defect information are realized.Through the construction of the online detection hardware system for detecting character defects,the research of character defect detection algorithm and the development of software system,the online detection and classification of printed character defects is completed.At the same time,the defect detection system is tested in terms of accuracy,detection accuracy and real-time performance.The test results show that the system can accurately and efficiently realize the detection and classification of printed character defects,and meet the requirements of printing character defect detection.