摘要
聚类是数据挖掘的重要分支之一,引入模糊理论的模糊聚类分析为显示数据提供了模糊处理能力,在许多领域被广泛应用。本文应用考虑邻域关系的约束模糊C均值(Fuzzy C-Means with Constrains,FCM_S)算法,将邻域像素引入到目标函数中,进而有效地利用邻域像素信息,提高分割精度。本文应用FCM_S算法对模拟彩色纹理图像进行分割,计算其混淆矩阵,定性定量地与FCM算法进行对比分析,证明了该算法的鲁棒性。
Clustering is an important branch of data mining , fuzzy theory of fuzzy clustering analysis provides a fuzzy display data pro-cessing capability , is widely used in many fields .In the paper , considering constraints neighborhood relations Fuzzy C -means ( Fuzz-y C-means with Constrains , FCM_S) algorithm will be introduced to the neighboring pixels objective function , thus the effective use of field-pixel information to improve classification accuracy .In this paper , FCM_S algorithm to simulate color texture image classifi-cation, calculate the confusion matrix , qualitative and quantitative comparison with FCM algorithm analysis to prove the robustness of algorithm.
出处
《测绘与空间地理信息》
2015年第6期98-100,共3页
Geomatics & Spatial Information Technology
关键词
图像分割
聚类算法
邻域约束
模糊C均值聚类算法
image segmentation
clustering
neighborhood distance constraint
fuzzy C-means clustering algorithm