Defect Detection And Research On Algorithm Of Printed Code On Bottle-Cap

In the process of Manufacturers use code printing machine printed code on the cap surface, due to imperfect code printing machines and printing materials, as well as some of the inevitable random factors, production date printed surface defects cap row codes, which may direct impact on the use of post-consumer goods, threatening the health of consumers, so the defect detection of prodauction date code on the cap is an essential part.In this paper, cap row code defect detection as the application background, combined with the actual production needs and the actual situation, the image pre-processing, image registration and defect detection algorithms research to identify suitable algorithm for this project, and through the experimental realization of defect detection code discharge cap.First introduced the standard template image preprocessing and production, through the acquisition of images of the noise analysis, using median filtering algorithm for image noise; In the standard image production session, through the analysis of two commonly used method of making a template, proposed a statistical method based on the average template image registration, avoiding image blur due to mechanical vibrations and the standard image made more clear and accurate. Secondly, the image registration algorithm conduct a study to ensure that the collected images to be detected, and the standard template images accurately registated, through in-depth analysis of the common image registration algorithm, we use the SURF-based image registration algorithm, better achieved with the image registration to be seized. Then, summarize some typical defect detection algorithm, based on the technical requirements, after analysis of poor video image binarization and de-noising method. Region growing connectivity and secondary connectivity analysis algorithms solve poor shadow image detection algorithm which can not have accurate statistics defect information and the noise sensitive issue. Finally, the image registration algorithms and defect detection algorithm proposed in this paper were cap row code defect detection experiments, analysis of the experimental results verify the feasibility of the algorithm.