摘要
该文给出了二维经验模态分解(BEMD)的实现方法,针对实现二维EMD的两种常用插值方式?三角剖分插值和径向基函数插值,通过理论和计算实践分析比较了两种插值方式的速度、视觉效果等,提出两种方法相结合的BEMD实现方法。同时,针对径向基函数插值方式提出了一种快速方法。该方法避免了求解大型线性方程组以及大矩阵的产生,较好地降低了时间和空间复杂度,提高了计算速度。实验结果证明了该方法的可行性和优越性。
In this paper, the implement process and application in image processing of bidimensinal ENID are given. Two interpolation methods are analyed and comperad in arithmetic speed and visual efficiency with experiments and theory, Delaunay triangulation based cubic spline interpolation and radial basis function interpolation which are usually used.At the same time, a new fast decomposition based on radial basis function interpolation is suggested. Its superiorities are also analyzed. The results prove this method is feasible, and its superiorities are obvrious.
出处
《电子与信息学报》
EI
CSCD
北大核心
2008年第12期2890-2893,共4页
Journal of Electronics & Information Technology
基金
黑龙江省教育厅科学技术研究项目(11511086)资助课题
关键词
图像处理
二维EMD
多尺度分析
径向基函数
Image Processing
Bidimensional Empirical Mode Decomposition（BEMD）
Multi-scale analysis
Radial basis function
