摘要
提出了基于水平集和迭代自组织数据分析的岩心砾石图像分割方法,既可克服水平集算法伴生的图像过分割问题,也可解决迭代自组织算法产生的图像边缘飘移问题。该方法首先运用水平集算法获取目标颗粒的准确边界,运用迭代自组织算法提取目标颗粒的整体;其次将两种方法提取的图像给以迭加综合处理,克服水平集算法伴生的图像过分割和迭代自组织算法产生的边缘飘移的缺陷,进而得到既完整又边界准确的目标颗粒。该方法运用于岩心砾石图像的分割,取得了良好的效果。
A method of gravel image segmentation based on leyel set and iterative self-organizing data analysis technique（ISO- DATA） is put forward to solve over-segmentation image made by Level Set and edge bias error produced by ISODATA.Firstly,it cuts out the precise edge of goal granule with Level Set,gets the wholeness of goal granule with ISODATA.Then,the images made by the two sets of algorithm are synthetically dealt with in a way of piling-up,in order to overcome the shortcomings existing in both Level Set and ISODATA,to get whole and edge-precise goal granule.It brings about a good effect in gravel image segmentation.
出处
《计算机工程与应用》
CSCD
北大核心
2007年第36期16-18,28,共4页
Computer Engineering and Applications
基金
四川省科技公关计划(the Key Technologies R&D Program of Sichuan(Province)
China under Grant No.05GG021-026-03)
关键词
水平集
图像过分割
迭代自组织数据分析
砾石图像
level set
image over-segmentation
iterative self-organizing data analysis technique
gravel image
