基于PCA的彩色印刷图像色彩分割算法

Color Segmentation of Color Printing Image Algorithm based on PCA

针对彩色图像的印刷过程中,原图像的色彩分割问题,提出了基于PCA(主成分分析)并结合其它典型彩色图像分割方法的新分割算法;该算法首先利用PCA算法把图像分解为主特征分量和残特征分量两分量图;然后采用二次分水岭算法对残特征分量图进行分割;利用K-Means算法对主特征分量图进行聚类初分割,接着对聚类初分割后的图像进行相似色彩区域融合;最后把分割后的两分量图的进行融合,得到最终的分割结果图;该算法可以应用于彩色印刷图像的色彩自动分割和彩色印刷过程的自动色彩控制中。

To solve the color image segmentation problem in the color image printing process,a new segmentation algorithm is proposed based on PCA and combined with other typical image segmentation algorithms.According to the proposed algorithm,firstly PCA is used to decompose the original image into two parts;secondly,the K-Means algorithm is used to divide the primal part of image and then twicewatershed algorithm is used to divide the remnant part of image;finally,the segmented images are got by merging the two kinds of segmentation images.This proposed algorithm can be applied to color printing image segmentation and automatic color printing control during the process of color image printing.