摘要
根据图像系统所固有的自相似性以及经验模式分解（EMD）算法的完备性和稳定性,给出了一种完整、快速、高效的EMD图像分解算法。该算法主要改进了EMD图像分解的曲面插值方法以及提取固有模式图像的结束条件,解决了一般EMD图像分解算法存在的图像分解算法速度慢、三角剖分导致漏点现象、图像分解算法的结束条件不明确的问题。最后,利用Matlab进行的仿真实验验证了该算法的有效性。
According to image systems＇ intrinsic quality of serf-comparability and the empirical mode decomposition （EMD） algorithms＇ completeness and stability, an improved algorithm for EMD image decomposition is p msentecl. It is integrative,fast and effective. It improved EMD decomposition＇ s bidimensional interpolation method and end conditions of getting intrinsic mode images, and solved the questions of general algorithms for EMD image decomposition, such as low speed in decomposition, failure to contain whole points, and the uncertainty of end requirements for image decomposition. The simulation experiments were made using Matlab and the effectiveness of the imprived algorithm was tested and verified.
出处
《高技术通讯》
CAS
CSCD
北大核心
2007年第11期1121-1124,共4页
Chinese High Technology Letters
基金
863计划（2002AA401001-4B）、国家自然科学基金（60475024）和航天技术创新基金（20060118）资助项目.
关键词
经验模式分解
图像分解
算法
empirical mode decomposition（EMD）, image decomposition, algorithm
