图像半色调化算法研究综述

Survey of Research on Image Halftone Algorithm

半色调化算法是图像打印输出的关键技术之一。介绍了简单阈值比较算法、使用随机数的半色调化算法、误差扩散算法、点扩散算法、按序抖动算法和抖动算法等一系列半色调化算法,并揭示了这些算法之间的联系。为了高效率地对图像做半色调化操作,商用光栅图像处理器一般使用抖动算法。抖动算法生成的半色调化图像的质量取决于所采用的抖动矩阵。基于几何分割设计半色调化抖动矩阵的方法可以在满足堆栈限制的前提下优化半色调化图像的质量。使用随机数的半色调化算法可以消除图像半色调化中产生的莫尔纹。随机半色调化算法的基础是将一定数量的随机点均匀地分布到一个正方形区域中。Asano提出了随机点均匀分布问题等价于圆形装填问题的命题,本文对该命题给出了初步的证明。最后,给出了一些有待进一步研究的图像半色调化课题。

The halftone algorithm is one of the key technologies for image printing. Introduced a series of halftone algorithms such as the thresholding value comparing algorithm,halftone algorithm with random thresholding values, error diffusion algorithm, dot diffusion algorithm, ordered dithering and dithering algorithm. Commercial raster image processor usually uses dithering algorithm for halftone speed. The quality of halftone image produced by dithering algorithm depends on dithering matrix. The geometry based method to design dithering matrix can optimize the halftone image quality while maintain the constraint of stack. The halftone algorithm with random thresholding values can eliminate the moiré of image. The foundation of random halftone algorithms is to distribute a number of points in a square area randomly and uniformly. Asano proposed a proposition that problem of distributing random points and circle packing in a square area are equal. This paper gives a basic proof to this proposition. Finally, a few topics regarding the to-be-investigated image halftone are proposed.