基于神经网络的智能油墨预置技术

Intelligent Ink Presetting Technique Based on Artificial Neural Networks

针对当前油墨预置技术存在的预置精度低且易受印刷参数和印刷条件影响的问题,提出了基于人工神经元网络的智能油墨预置新技术.以实际合格印品为学习样本,将印刷条件参数和网点面积率参数作为输入特征量,实际的墨键开度值作为网络输出,选择3层BP网络构建印品图文数字信息与墨键控制参数间的映射关系,实现数字化印刷流程中快速、准确的油墨预置.多色胶印机印刷试验结果表明,该方法能有效缩短开机调整时间,提高印刷效率,降低生产成本.

In order to improve the ink-presetting precision in the different printing parameters and printing environment,in this paper a novel approach using ANN technique is presented.For precision and rapid ink-presetting in the digital printing process,with the actual qualified printing products as training samples,the printing parameters and dot area percentage as input characteristic values and actual ink key values as ANN output,a three-layer BP network is selected to establish the mapping between the graph-text information and the values of the ink key.The experiment result shows that this method can shorten the adjusting time of the offset press effectively,improve the printing efficiency and reduce production cost.