摘要
针对基于势函数聚类的多阈值图像分割算法的不足,定义了伪势的概念,并在原算法基础上提出了一种改进的图像分割算法。由伪势概念确定了伪势合并的判别方法,按照此方法,当相邻的两个峰之间的距离小于所定义的自适应模糊伪势因子时,则应该进行伪势合并。改进后的算法在计算剩余势函数时判断是否存在伪势,然后在势划分函数组的确定过程中相应地进行伪势合并计算。利用多幅图像进行了多阈值分割的仿真试验,结果表明,改进的基于势函数的多阈值图像分割算法具有更好的鲁棒性和分割效果。
Aiming at the disadvantage of multi-threshold image segmentation algorithm based on potential function clustering, a concept of false potential is defined and an improved image segmentation algorithm is proposed on the basis of an original algorithm. According to this method, when the distance between the adjacent peaks is less than the adaptive fuzzy false potential factor, the false potential mergence should be implemented. The improved algorithm will judge if there are false potentials when calculating the residual potential function and if false potentials exist, implement false potential mergence in the process of dividing the potential function groups. The segmentation experiment for several images shows that the improved algorithm possesses better segmentation and more robust performance than the original algorithm.
出处
《光电工程》
EI
CAS
CSCD
北大核心
2005年第8期64-68,共5页
Opto-Electronic Engineering
基金
中国科学院科技创新基金资助项目(A010416)
关键词
图像分割
势函数
多阂值
鲁棒性
Image segmentation
Potential function
Multi-threshold
Robustness
