基于均匀色差的彩色图像分割方法  

Color Image Segmentation Based on Average Color-Difference

作　　者：王小青[1] 位军[2] 陈素军[3] 

机构地区：[1]中国兵器工业第五八研究所军品部,四川绵阳621000 [2]西北工业大学电子信息学院,陕西西安710072 [3]燕山大学信息工程学院,河北秦皇岛066000

出　　处：《兵工自动化》2006年第10期42-43,46,共3页Ordnance Industry Automation

摘　　要：基于均匀色差的彩色图像分割方法,结合模式识别理论的自动聚类彩色图像分割。其图像首先被转化到L*a*b*颜色空间,进行初始化并确定聚类中心。然后计算像素到聚类中心的色差,进行聚类。之后计算新的聚类中心,反复迭代,直到符合终止条件,得到最终分割结果。实验证明,该方法可很好实现彩色图像的自动分割。Based on the color-difference, the color image segmentation combined with the automatic clustering color image segment of mode identification theory. Its images were transformed into the L^＊a^＊b^＊ color space at first, and carried out the initialization and ensured the clustering center. Then, the color-differences between picture elements and clustering center was calculated and carried out the clustering. After that, the new clustering center was calculated and iterated, till they fitted the stop conditions and divided the results at last. The results showed that this approach could realize the auto segment of color image effectively.

关 键 词：均匀色差 彩色图像分割 L^＊a^＊b^＊颜色空间 聚类中心 

分 类 号：TN911.73[电子电信—通信与信息系统]