欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
针对Krinidis等人提出的模糊局部C-均值聚类系列算法缺乏合理性的不足，提出一种新的鲁棒模糊局部C-均值聚类分割算法。对鲁棒模糊局部C-均值聚类的目标函数重新分析并建立正确的聚类目标函数，对新的聚类目标函数及其约束条件采用拉格朗日乘子法进行严格数学推导并获得一种新的隶属度和聚类中心迭代表达式，最后设计一种新的充分利用像素邻域信息图像聚类分割算法。实验结果表明，所建议的鲁棒模糊局部C-均值聚类分割算法是有效的，相比现有鲁棒模糊局部C-均值聚类分割算法更适合复杂遥感等图像分割的需要。
英文摘要：
Aiming at the fuzzy local C-means clustering series algorithm proposed by Krinidis et al is lack of rationality, this paper proposed a new robust fuzzy C-means clustering segmentation algorithm based on local information. It firstly reanalysed the objective function of robust fuzzy local C-means clustering and established the right clustering objective function. Then the new clustering objective function combined with the constraint condition of fuzzy partition membership was strictly mathematical deducted to obtain new iterative expressions of membership degree and clustering centers by Lagrange multipliers. In the end, it designed a new image clustering segmentation algorithm by utilizing pixel neighborhood information completely. Experimental results show that the proposed robust fuzzy local C-means clustering segmentation algorithm is effective, and more suitable to segment complex remote sensing image than the existing robust fuzzy local C-means clustering segmentation algorithm.
同期刊论文项目
同项目期刊论文
期刊信息
位置：