摘要
读入印前数字文件,计算每一墨区的着墨面积比例,利用人工神经网络算法生成一个能反映印刷机印刷当前版面时的油墨传递特性的矩阵。当给定纸上墨量要求时,利用奇异值分解法求解方程组,通过截断小的奇异值的方法,求取一种墨量预设值,在该预设值条件下,承印物上的墨量虽然不能均匀一致,但偏差均方根小于规定的要求。基于上述研究,给出一种实现平版印刷墨量预设的新方法,经实验和生产实践检验,该方法具有较高的准确性。
After reading the digital prepress file, the inking area ratio of each ink zone is calculated. Artificial neural network algorithm is used to generate a matrix that can reflect the ink transfer characteristics of the press in printing current document layout. When the ink quantity requirement on paper is given, matrix singular value decomposition method is used to solve the equation system and obtain the ink preset value with the truncated singular value method. Under this ink preset value, although the ink quantity of every zone on the substrate is not uniform, but the root mean square error is less than the predefined value. On the basis of above study, this paper proposes a new method for ink preset in offset printing. Experiment and practical production verify that the proposed method has good accuracy.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2013年第6期1353-1358,共6页
Chinese Journal of Scientific Instrument
基金
浙江省自然科学基金(Y1100771)
浙江省教育厅项目(Z201120502)
绍兴市科技计划(2012711)资助项目
关键词
平版印刷
油墨传输
奇异值分解
油墨预置
offset printing
ink transfer
singular value decomposition
ink preset