中文摘要：

针对相似图像分割过程中,输入像素数据在转换空间上存在的不连贯和幅度变化特征差异很小,像素的隶属关系很难准确界定,导致分割阀值设定过程出现较大衰减,分割误差较大的问题,提出一种改进的模糊聚类图像分割算法。分析了传统的模糊C-均值聚类图像分割算法的弊端,对像素模糊划分矩阵和聚类中心进行推导,将迭代过程中像素数据集对聚类隶属的可能性和不确定性关系融入分割目标函数中,依据可能隶属度和不确定隶属度建立改进分割准则函数,同时对像素聚类进行更新,实现图像分割。仿真结果验证了所提算法的有效性,结果表明,改进后的方法在分割检测过程中,图像误差明显减小。

英文摘要：

An improved fuzzy clustering image segmentation algorithm was put forward to solve the problems of bigger attenuation of the segmentation threshold value setting process due to the incoherence of data transition space of input pixel data in the similar image segmentation process, diminishing amplitude of characteristic difference and the difficulty of accurate definition of the pixel subordination. The disadvantages of the traditional fuzzy C - average clustering image segmentation algorithm were analyzed, and the pixels fuzzy partition matrix and cluster centers were derived. In the iterative process, the uncertainty and possibility of the relationship of cluster to pixel data set were integrated into the segmentation in the objective function. According to the possible membership degree and uncertainty membership degree, a criterion function of improved segmentation was established, and the segmentation of pixel clustering was updated at the same time, so as to complete the image segmentation. The simulation results verify the effectiveness of the proposed algorithm, and the improved method in the segmentation process can reduce errors significantly.