Research On Visual Inspection Technology Of Surface Printing Defects In Wine Bottles

With the continuous improvement of production and technology,the printing of bottle labels tends to be automated.The quality of label printing is a concern for most bottle manufacturers.The bottle label printing defects are unprinted,missing prints,incomplete,scratches,blisters,pasteboards,etc.Currently,the detection of wine bottle label printing quality by domestic companies is mainly a traditional manual visual inspection method.This detection method is slow.,Manual detection of fatigue,the existence of subjective factors affect the testing standards,it is difficult to meet the requirements of high-speed automated production lines.Therefore,it is of great significance to study the defect detection system based on machine vision to improve the efficiency of automated production.This article introduced the key technology of the visual inspection of the bottle label printing defects.According to the characteristics of the surface label of the bottle,the high-precision imaging system of the label printing on the surface of the online bottle was designed,and the related algorithm of defect detection was studied in depth.A visual inspection software system for surface defects of wine bottles.The main work of this paper includes the following aspects:1.According to the characteristics of the bottle printing labels and the defect detection requirements,a visual inspection system for bottle label printing defects was designed.In terms of imaging,a combination of line scan cameras combined with a rotating platform and backlighting were used to achieve good imaging results.Visual inspection combined with mechanical structure for automatic loading and unloading to achieve automatic defect detection on the assembly line.2.The type of defect was analyzed.Firstly,the image was tested for the integrity of printing,and whether the bottom oil,patterns and fonts were printed.For the unprinted condition,preliminary detection is performed by using the gray value feature judgment method;the vertical projection is used to find the edge to locate the edge of the pattern;when the bottom oil is detected,the method of image enhancement and vertical projection detection of the edge pair is well solved.The problem of grayscale is not obvious.3.In the defect detection of the pattern,the method of comparing the image to be detected with the standard template is used.In this paper,shape-based template matching is used to solve the problem of partial pattern missing affecting the matching result.The method of image pyramid pyramid detection,limited search area,and multiple template combination method is used to improve the detection speed and accuracy.A block registration method is proposed to solve the problem of different starting points for image acquisition,which effectively reduces the registration error.In order to remove the difference in gray levels of back images and labels,binarization is performed before the registration images are differentiated.4.Defect analysis:In order to remove the interference and error information at the back edge of the difference,morphological processing was used to deinterfere the difference image.Finally,the Blob analysis feature was used to analyze the processing results,and the Blob analysis results of multiple blocks were integrated to realize the entire tag Defect detection.