Research Of Print Quality On-line Detection Technology Based On Visual Characteristic

Printing quality detection is a key link in the printing process, it relates to the printingquality and production efficiency. Contraposing traditional printing quality inspection notdetect printing quality based on full picture in the real-time, and also hard to reflect the realproblem of printing quality characteristics, we put forward a system for printing qualityevaluation，which is based on visual features，and study the printing quality online detectiontechnology to improve the accuracy of printing quality detection.This paper did research mainly in four aspects，which are printing quality evaluationsystem, text printing quality evaluation, image printing quality evaluation and printingquality online detection scheme. Firstly, analyze and summary the traditional printing qualityevaluation system and its shortcoming. According to human visual characteristics and printingquality evaluation index, using the strategy of “region of interest” in visual, construct theprinting quality evaluation system. Secondly, text printing quality evaluation templates wereestablished by word binarization and refinement, and using structure similarity algorithm torealized quality evaluation of printing text area. Then, color printing image was turned intoquaternion form, and using quaternion singular value decomposition strategy to evaluated thequality of colorful printing image from the image structure. Finally, printing quality onlinedetection scheme was established. In the scheme, applied inverse halftoning processing andimage position calibration to the printing image, and used quality evaluation method of textand image to realize printing quality online detection according to the printing qualitydetection standard. The experimental result shows that this method realize the effectivecombination of subjective evaluation and objective evaluation of the printing quality, andensure the accuracy of the online printing quality detection, make evaluation result consistentwith the human visual characteristics.This paper studies the related technologies of the online printing quality detection basedon visual characteristic, and acquired a printing image quality evaluation method which wasin accordance with the visual characteristics, to offset the problems that printing qualityevaluation results were not in accordance with the visual characteristics. This method madethe printing quality detection more accurate, and laid a technical foundation for keyalgorithms of the printing image quality evaluation based on subjective and objectiveinformation fusion.