基于空间信息的直觉模糊C-均值图像分割算法

针对模糊C-均值聚类算法的单一隶属度不能充分描述图像不确定性,且聚类过程中忽略像素空间关系的问题,提出一种基于空间信息的直觉模糊C-均值算法；该算法选取3×3的模板计算邻域像素灰度均值；并引入权重项,来控制灰度信息和空间信息各自所占的比重,同时用犹豫度更新直觉模糊集的隶属度函数；对常用标准图像的仿真结果表明,该算法能更好地保留图像细节信息,得到更加理想的图像分割效果。

Image Segmentation Algorithm Based on Spatial Information of Intuitionstic Fuzzy C-Means

In view of the Fuzzy C-Means clustering algorithm of the single uncertainty membership degree can not fully describe the images, and ignore the pixel spatial relations in the process of clustering, here put forward a kind of image segmentation algorithm based on spatial information and intuitionistic fuzzy sets.The algorithm select the template of 3 3 computing neighborhood pixels within the grayscale average; and introducing penalty term for the fusion of gray level information and the grey value of pixel information space, at the same time introducing spatial neighborhood information in the objective function. Select the commonly used standard image to segment, the experimental results show that the intuitionistic fuzzy C-Means algorithm based on spatial information has strong resistance to noise, at the same time also can keep image details and obtain a more ideal image segmentation results.