摘要
在定义乘性可变尺度结构算子对数学形态维分形模型进行优化和改进的基础上,提出了一种基于提取分形特征参数的图像分割算法。首先定义并比较了加性和乘性可变尺度形态结构算子,并以此为基础进行形态学膨胀运算,提取归一化模型下的数学形态维分形维数,再用分形特征维数调制灰度值,以拉开各个灰度级间的距离,有效地增大了目标和背景的差异,最后进行自适应阈值分割。大量对比仿真实验取得了良好的分割效果,并且证明了本算法的有效性和可靠性。
Based on defining variable structure elements to optimize fractal morphology model, a fractal dimension segmentation algorithm was proposed to get better results. At first, addition and multiple structure elements were defined and compared to fit in morphology operations. Then, fractal morphology dimension was abstracted and used to modulate gray scales, in order to increase contrast of target and background. At last, a threshold was automatically determined for segmentation. Abundance experiment data support its availability and credibility.
出处
《计算机应用研究》
CSCD
北大核心
2007年第4期151-153,共3页
Application Research of Computers
基金
国家"863"计划资助项目(2003AA823050)
关键词
数学形态维
归一化
分形维数
阈值分割
分形调制
fraetal
morphology dimension
normalized
fractal dimension
threshold segmentation
fractal modulation
