具强噪声鲁棒性的聚类图像分割算法

On Strong Robust Clustering Image Segmentation Approach

针对已有的图像分割算法对噪声的鲁棒性不佳的问题,结合中智集合提出了一种新的基于中智集合聚类的图像分割算法,并为此聚类算法提出了高性能的目标函数.首先,将图像转换为中智集合域,然后定义一个高效的基于中智集合的聚类目标函数进行聚类分析,最终采用聚类算法将像素进行分类.将人工图像与真实图像进行对比试验.结果证明:本算法的有效性与分割准确率均高,同时具有较好的噪声鲁棒性.

Aimed at the bad robust feature of the existing image segmentation approach,a new segmentation approach has been proposed by combining with the neutrosophic set clustering,and an efficient object function been proposed for it.Firstly,the image is convertedto neutrosophic set domain;secondly,an efficient object function has been defined and clustering been analyzed;at last,the pixels have been clustered by the clustering approach.Experimental results prove the efficiency and accuracy of the approach,which has strong robust performance.