AbstractsComputer Science

Flexographic print is a technology that is growing in popularity, especially in the packaging industry,due to its ability of printing on a wide variety of substrate types. Advances in ensuring good quality from flexographic processes have helped to bolster flexographic printing’s limited usage in the past decades. However, improving the quality of flexographic print products to a level that matches that of other mainstream printing methods is still needed. One common defect that affects flexographic print quality is mottling, which has been studied extensively in e.g. offset prints. At Innventia, a dependable method of calculating mottle numerically that highly correlates with subjective observations has been researched. This method has been applied into developing software called STFI-Mottle, which processes scanned images of halftones and returns average mottle values for them. In the case of this thesis, the goal has been to investigate what sort of characteristics in halftones caused by flexographiclead to a diminished halftone print quality on coated kraft board substrates. These characteristics were studied by examining dot uniformity in pre-printed samples of 30% halftone prints, then evaluating the data gathered with the ambition of finding a correlation between dot feature variations and print mottle. The mottle values that were used for analysis and discussion were computed using STFI-Mottle.The initial reasoning was that “white spots” found within the samples’ raster dots should be closely examined, as they might provide clues to explaining unexpected print mottle. The 30% halftone cyan regions of printed samples were scanned with a high-resolution Creo iQsmart3uniformity analysis was done with Matlab-based dot property calculation software. Thresholding was employed in image analysis in order to separate printed dots from unprinted substrate surface. A mottle evaluation was performed, focusing on the mottle values that were measured in the R channel using the same RGB images that were produced during scanning. Every data variable was compiled into a mean average for each sample, based on five to three signatures. Distribution histograms, simple linear regression and multiple linear regression were used to find meaningful correlations between measured data and mottle values.The study resulted in some significant correlations between variables established as quality indicators, but also showed possible new connections between variables that can express dot uniformity and print mottle. Some variables’ viability as mottle predictors has been rejected, while others are presented as new opportunities for study. The deeper implications of some of the variables analysed could not be investigated, but a foundation for a methodology to conduct similar research has been laid.Furthermore, the thesis has yielded an initial base of knowledge on new objective data that might be used as alternative measurements for quantifying print quality. In conclusion, the study indicated that there are dot uniformity variables…