摘要
针对特征颜色提取这一陶瓷砖色差分级的关键问题,提出了应用灰色系统理论的改进特征聚类算法。首先在将图像转换到HSV颜色空间,并根据实际情况确定初始聚类中心,然后计算图像中每个像素点到各聚类中心的灰色关联度,最后应用改进的聚类运算完成特征颜色的提取。对不同纹理的陶瓷砖进行了特征颜色提取实验。实验结果表明,改进算法可准确有效的提取较为复杂的随机纹理陶瓷砖的特征颜色,与人的视觉判断相一致,且在计算效率方面比传统的模糊聚类算法有较大提高。
A feature color extracted algorithm based on the gray relational grade is proposed to solve the problem of detecting the color difference of ceramic tiles.Firstly,the number of clusters are calculated and the initialcluster center is determined in the HSV color space.Then,the gray relational grade of pixel is calculated and the feature color is extracted with the improved gray clustering algorithm.Taking the ceramic tiles for examples,the experimental results prove that the algorithm can be used to extract the feature color of the random texture.
出处
《中国陶瓷》
CAS
CSCD
北大核心
2012年第12期58-60,93,共4页
China Ceramics
基金
河北省科学技术研究与发展计划项目(11215640)
唐山市科学技术研究与发展计划项目(11110210B-6-6)
关键词
特征颜色
灰色关联度
灰色聚类
算法
Feature Color
Gray Relational Grade
Gray Clustering
Algorithm
