摘要
基于轮廓序列的三维重建需要解决不同层面的匹配与聚集问题,传统算法在复杂流型转换中匹配准确率较低。针对传统匹配算法匹配准确率较低等问题,采用对折线求平均距离、构造模糊集的方法,给出了一种模糊聚类的数学模型,将传统的匹配问题转化为类成员的隶属度问题,最后把该模型应用到电容层析成像三维可视化系统中,并通过仿真实验验证了方法的有效性。
In 3D reconstruction of contour sequence, the matching and clustering in different layers should be firstly resolved. However, in processing the complicated flow pattern transfer, the accuracy rate of traditional algorithm was considered very low. Based on the above-mentioned problems, a mathematic model of fuzzy clustering by computing the average distance between two poly lines and constructing fuzzy sets was established. The problem of matching was transformed into the grade of membership of elements by this mathematical model. Finally, this model was applied to the system of ECT 3D visualization, and the validity and reliability were tested.
出处
《系统仿真学报》
CAS
CSCD
北大核心
2008年第15期4078-4081,4095,共5页
Journal of System Simulation
基金
国家自然科学基金(60572153)
国家教育部重点科技项目(204043)
黑龙江省重点科技攻关项目(GC05A510)
黑龙江省自然科学基金(F200609)
哈尔滨市重点科技攻关项目(2005AA1CG035)
关键词
轮廓匹配
隶属度
聚类
三维重建
contour matching
grade of membership
fuzzy cluster
3D reconstruction
