基于K-means聚类的纺织品印花图像区域分割

Research on segmenting color regions in textile printing image based on K-means clustering

对纺织品彩色印花图像进行颜色区域分割.将彩色纺织品印花图像转换到CIEL*a*b颜色空间,用K均值聚类分析算法对描述颜色的a*和b*通道进行聚类分析;通过提取各个颜色区域独立成为单色的新图像,对彩色纺织品印花图像进行分割处理.实验结果表明,在CIEL*a*b空间使用K-means聚类算法可以有效地分割彩色纺织品图像的颜色区域.

Image segmentation in textile printing image was realized. The textile printing color image was converted to L ＊ a ＊ b color space. The a ＊ and b ＊ are selected and clustered by K-means clustering algorithm. After extracting each color region and create images that segment the textile printing image by color, the segmentation was realized. The experience result proves that color-based segmentation can distinguish the textile printing image effectively.