目的 借助数字图像处理方法提取网点结构， 直观反映网点的微观形态特征。方法 针对光学显微镜采集的网点图像， 提出一种利用中值滤波去除网点图像噪声， 使用非极大值抑制原理判断边缘点， 结合二值化和细化， 对网点结构进行最后处理的微观结构提取方法， 寻找最真实的网点与承印物边界； 通过模拟测试图， 验证了提出方法的合理性。结果 使用该方法得到的网点结构图效果明显优于传统边缘检测方法， 能有效抑制背景噪声， 较大程度地保留非重叠网点边缘信息， 且客观评价指标良好。结论 提出的算法为基于微观网点评价印刷品质量提供了理论依据。

The aim of this study was to reflect micro-structure characteristics of dots by means of image processing method for extracting dot structure. Specific to the dot image collected by optical microscope, a method of extracting micro-structure of dots was proposed. In this method， median filter was used to remove the image noise， the principle of maximum inhibition was adopted to judge the edge point， binarization and thinning were used in combination for the final process of the dot structure to find out the realist boundary of dots and printing stocks. In this paper，the simulation testing figure was used to verify the rationality of this method. The dot′s structure by using this method was better than that obtained by traditional edge detection. This method can effectively inhibit background noises, especially for non-overlapping dots whose edge information was retained to the largest degree. And the objective indexes were evaluated as good. The proposed algorithm based on the micro-dots provides a theoretical basis for evaluation of printing quality.