摘要
传统的分水岭分割算法受噪声和图像细节信息的影响,存在过分割、对噪声敏感等缺陷.针对这些缺陷,提出一种基于密度模糊聚类的分水岭分割算法.首先对图像进行分水岭分割,提取各子区域灰度均值,然后对灰度均值进行密度模糊聚类,进行区域合并.进行多组对比实验,结果表明此算法具有可行性和有效性.
Because of the noise and the influence of the information of image details, the shotcoming of the traditional watershed algorithm lies in much segmentation, sensitive to noise and so on. Therefore, the paper comes up with watershed segmentation algorithm based on fuzzy density clustering. First image was segmented by watershed, the small area average value was extracted, then the image for gray average value was merged by fuzzy density clustoring. Many contrast experimental results verified the feasibility and validity of the method.
出处
《安徽大学学报（自然科学版）》
CAS
北大核心
2013年第3期56-60,共5页
Journal of Anhui University(Natural Science Edition)
基金
安徽省教育厅自然科学基金资助项目(KJ2007B069)
安徽大学"211工程"学术创新团队基金资助项目(KJTD007A)
关键词
分水岭
模糊密度聚类
形态学重建
点密度
区域合并
watershed segmentation
fuzzy density clustering
morphological reconstruction
dotdensity
area merge
