【摘要】 图像陷印技术是印前处理的关键技术之一,通常的图像陷印都是基于邻域分析的,不能够从整体上反映一些大尺度趋势,对图像的破坏较大,并且不便于对处理结果进行修改和编辑。本文提出一种基于边界的图像陷印方法,能够从大尺度上反映图像内容的变化趋势,得到稳定直观的陷印效果,并且方便人工事后编辑。该方法主要包括以下几个步骤:首先通过彩色边缘检测算法得到边界的精准位置和描述,并且计算出边界两侧的像素点与边界的距离关系;其次,分析边界两侧的颜色信息获取边界两侧色块的表征性颜色;然后,根据边界两侧的表征性颜色和陷印参数在边界距离关系的基础上进行整体性的陷印。由于记录了边界的信息及边界距离关系信息,因而能实现中分,自然连续等多种陷印要求,并且可以方便的实现以边界为整体对陷印结果进行编辑。实验表明该方法具有邻域分析方法无可比拟的优越性。更多还原

【Abstract】 Image trapping is a very important technology in pre-printing.Traditional methods tend to perform trapping by neighborhood analysis,which is short of presenting the global trend in a large scale and has obvious degradation effect on image quality.What’s more,modification of trapping result of neighborhood method is very difficult.In this paper,a new image trapping method based on edge analysis is proposed.This new method can solve the above shortages of neighborhood method.It is composed of three steps.First step is to precisely detect color image edge and calculate edge distance map.Second step is to analyze the color information of contents on both side of the edge.The last step is to perform trapping based on color information of each connected edge chain achieved in the first two steps.By presented method,"trapping into middle"and"natural continuous trapping"is made possible. Trapping results based on edge analysis can also be easily modified as a whole.Experiments show that the new method has distinct advantage than traditional neighborhood methods.更多还原