A Study On The System For Dot Acquisition And Separation In Color Printing

With the tendency of the printing to low-volume, multicolor, high efficiency, it is inevitable that online-detection & control of the presswork color quality come forth. Traditional detection on the ground of density and chroma is based on color-mixing models. Researches relevant to this have already recovered the disadvantages of this theory, and the detection technology based upon this can only detect the accessional especial signal trip in the presswork. A new detection technique needs to be advanced to detect the dots of printing image directly.In order to detect various dots of color printing image, hardware system of acquisition and separation in microscopic dot image is designed. The system includes three parts: image acquisition, image processing and function control. Microscopic dot image is captured by image acquisition. Image processing based on digital image processing and pattern recognition finishes dot separation and offers the result. Function control controls image acquisition and processing. Aimed at the color trapping of color microscopic dots images, the algorithm is designed on the basis of the concepts of Skiving-Rind. Clustering, decision tree, neural networks, fuzzy logic and information fusion etc. are been used in classifiers, which make the classifiers can meet the real time of online-detection, and also have property of local optimal approximation. In the end, the classifiers are verified by 16- and 24-bit images and the error analysis is given. The next research direction is pointed out.The thesis will not only practical in online-detection & control of presswork color quality, but also contribute a lot in lucubrating color dots color theory and printing ink superimposition rule.