Study On Hysteresis Coefficient Of Digital Screening

Digital printing has developed rapidly in the recent years, but its printing quality and speed can not meet people's demand. The error diffusion with output-dependent feedback (EDODF) is the core of the popular amplitude modulation (AM) and frequency modulation (FM) screening technologies nowadays. And the dot clusters is adjusted by Hysteresis Coefficient (H). Many researches about the algorithm are going with certain constant H, but the essential problem is still not solved. The object of this thesis was the screening technology of digital printing, and the selecting rule of the H was acquired. The purpose was to make the H more accurately, improve the dot output precision, obtain the high quality digital printing products, satisfy the people with the continuously rising demands of matters and culture, reduce the cost and save the energy and materials. The conclusions might provide theory and practice basis for the popularization of digital printing, and also had important significance for promoting the development of digital printing technology.The study methods: Based on the image process theory and the digital screening principle, according to the numerical scope limit, the input coefficient grayscale value of feedback equation was reduced to attain the balance of grayscale-output The best Hysteresis filter was chose by the main capability of filters. The optimal H for the typical grayscale progression was filtrated out by the theory that the bigger the PSNR, the more similar between the processed image and the original image. This H was corresponding to the maximum PSNR. And the optimal Hysteresis coefficient of the typical grayscale progression was processed with the least square method, and the curve equation of H was got The dot gain was computed and evaluated by Murray-Davis equation.The main works: Input coefficient of feedback equation was confirmed by experiment The grayscale progression output of the EDODF was limited to normal range to eliminate grayscale error. In the computer experiment, the Floyd-Steinberg was chose by comparing the capabilities of some hysteresis filter in common use, including the process speed, quality of screening algorithm and dot cluster of dot congregating when H became bigger. Valid range of H was obtained according to the PSNR of the standard image, which was screened with different H. The H value for each progression was filtrated through processing gray gradation using different H value. The relation curve between the optimal Hysteresis coefficient H and grayscale progression was obtained by fitting. The image was screened by different error diffusion algorithm, and then printed. After that the dot gain of output proof and image detail reproduction were evaluated synthetically.The main conclusions: After the input coefficient of feedback equation (2-8) was modified, grayscale value output scope was limited within [0,255]. Grayscale error in the primary equation was eliminated. The results of screening experiment showed that the Floyd-Steinberg filter was better man others. The bigger H, the more thickly the dot congregated. The relation between Hysteresis coefficient and grayscale progression was non-linear, and its function was confirmed. According to current pixel grayscale value, the best Hysteresis coefficient could be gained. Consequently, different dot clusters at different grayscale progression was gained. EDODF with Hysteresis coefficient could bring the better quality of image detail and more accurately reproduction of original picture. The dot gain phenomenon was controlled effectively by congregating the random dot to dot cluster.