基于图像多目标分割的水墨画风格仿真方法研究  

Research on Simulation Method of Ink Painting Style Based on Multi-objective Image Segmentation

作　　者：李博[1] LI Bo(Shaanxi Railway Institute,Weinan 714000 China)

机构地区：[1]陕西铁路工程职业技术学院,陕西渭南714000

出　　处：《自动化技术与应用》2021年第10期81-85,共5页Techniques of Automation and Applications

基　　金：陕西省教育厅人文社会科学项目(编号18JK0327)。

摘　　要：针对传统的多目标分割算法在进行中国水墨画图像分割仿真过程中渲染效果不佳,细节处理粗糙的问题。本文通过数字图像处理方法,根据先验知识将相关物体关联,利用二维多目标图像分割技术寻找分割对象边缘,引入相似度准则获取种子像素生长的起点,采用颜色匹配距离进行相似度区分,有效利用图像空间信息来提升对图像颜色的匹配。通过将本文改进的算法与传统阈值法进行实例验证表明,本文提出的分割算法获得的水墨画仿真图片具备了优异的图画分割渲染,大幅提高了水墨画的细节处理效果的同时,具备了高的运行时效性。As traditional multi-objective segmentation algorithm in the Chinese ink painting image segmentation simulation process rendering effect is not good and has the details of rough processing,a method is proposed,the related objects are associat-ed according to the prior knowledge by digital image processing method.The edge of segmentation object is found by us-ing two-dimensional multi-objective image segmentation technology,and the starting point of seed pixel growth is ob-tained by introducing similarity criterion,and then the color matching distance is used to distinguish the similarity degree.Image spatial information is applied effectively to enhance image color matching.Experimental verification results of the improved algorithm and traditional threshold method show that the simulation picture of ink painting obtained by the pro-posed segmentation algorithm with excellent picture segmentation rendering has high operational timeliness,which great-ly improves the ink painting detail processing effect.

关 键 词：多目标分割算法 水墨画 相似度准则 图形渲染 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]