数字图像边缘检测的模糊聚类研究

Edge Detection of Digital lmage through Fuzzy Clustering

数字图像的边缘检测实际上是一个聚类的问题，这些边缘由许多线段、曲线或平面和曲面所构成，组成了不同的类属和分割的区域，可以用在机器人视觉的模型识别、军事目标识别等领域．模糊聚类法（ 包括模糊均值聚类法和模糊球壳聚类法） 具有不必事先对数据中的边界特征信息了解太清楚的优点。

Edge detection in digital images is a clustering problem for these edges are composed of lines and curves, or planes and arbitary shapes which belong to different clusters and areas. In this paper,Fuzzy c-means and fuzzy c-shells algorithms are presented to process iris data and baboon image together with edge detection methodologies. Its results imply that the proposed algorithms have advantages of capability of process without prior-knowledge of original data and image and can be used in robot vision, military target detection, etc.