Research Of Dot Identification And Detection In Full-Screen Detection

In this paper, major research direction is process of dot percentage, through analysis Plane model, Neugebauer equations, Artificial neural network and so on, the paper bring up a kind of dot percentage solving method based on pattern recognition of distance discriminant function. Experiment procedure start by the following aspects:The first is a experiment sample selection, this article select746shades of gray in IT8.7/3serve as sampled data, acquire Lab and CMYK value of746shades of gray, at the same time analyze Color Space Theory, establish pixel by CCD collected transformational relation from RGB Color Space to CIE-L*a*b*Color Space. Secondly set up discriminant function, according to16kinds of color modes of746shades of gray, apply MATLAB software and matrix equation, obtain16kinds of distance discriminant function and Lab color feature standard value of each color mode, in this way we can detect color modes of pixels or blocks of color under test. Once again calculate dot percentage and write computer programs, calculation the ink zone dot percentage by each pixel color mode in the ink zone, mainly apply the theory is C,M,Y,K single number divide CMYK total quantity, at the same time write all the steps into to the computer programs by C++language. In the end design verification edition and check accuracy of pattern recognition, design verification edition of16kinds of color modes, Proof,output and measure data, verify by computer program, at the same time analysis the accuracy and draw a conclusion, for monochrome, bicolor and not contain K in three-color detection accuracy is pretty accurate, but for contain K in three-color and four-color detection accuracy is not very well, the main reason is color gamut range is too small of the experimental sample, it can not contain color gamut range of the print.Pattern recognition of distance discriminant function bring up a new idea for dot percentage in ink zone, through the analysis of experimental error can know accuracy primarily depends on the selection of experimental samples, the experimental samples must have a enough large color gamut, in this way pattern recognition of distance discriminant function solve dot percentage will greatly improve the accuracy, the conclusions of this paper also set a direction for the follow-up studies.