基于单位纯四元数空间上聚类的彩色图像分割研究

Color Image Clutering Segmentation Based on Pure Unit of Quaternion Space

在RGB空间中,通过四元数模型把彩色图像像素三个分量整体进行处理,利用单位纯四元数空间中三个四元数乘积的性质,按照k均值分割算法思想建立单位四元数空间的彩色图像分割Q-Kmeans算法。实验结果表明,该算法具有良好的分割性能和实现上简单、速度快等特性。

This paper presents a new color image segmentation approach representing clustering algorithm based on pure unit of quaternion space. The clustering algorithm is the name of Q - Kmeans algorithm which regards the three components of color image by quatemion means as a whole, and analysis quaternion product result by K - means clustering. The experimental result shows that the Q - kmenas algorithm has good characteristics of high efficiency and fast speed, etc.