摘要
针对二维局部均值分解(BLMD)对图像进行多尺度分解时,计算过程费时,模态混叠现象严重的问题,提出一种快速自适应二维局部均值分解算法(FABLMD)。算法首先通过图像极值点信息计算顺序统计滤波器(OSF)窗口大小;然后利用OSF求图像上下包络曲面,进而得到包络估计函数和局部均值函数;最后在新的筛分终止条件下经过有限次迭代,快速将图像分解成一系列尺度不同的分量图像。仿真结果表明,改进算法计算速度更快,模态混叠现象得到有效抑制,对图像中不同频率成分的分解能力和效率都要优于BLMD。
A fast adaptive two-dimensional local mean decomposition algorithm(fablmd) is proposed to solve the problems of time-consuming calculation process and serious modal aliasing in multi-scale image decomposition by two-dimensional local mean decomposition(blmd). The algorithm first calculates the window size of the order statistical filter(OSF) based on the image extreme point information;then uses the order statistical filter to find the upper and lower envelope surfaces of the image, and next obtains the envelope estimation function and the local mean function;Finally, after a limited number of iterations under the new screening termination condition, the image is quickly decomposed into a series of component images with different scales. The simulation results show that the improved algorithm has faster calculation speed, modal aliasing is effectively suppressed, and the decomposition ability and efficiency of different frequency components in the image are better than BLMD.
作者
马朝永
付云超
胥永刚
MA Chao-yong;FU Yun-chao;XU Yong-gang(College of Mechanical Engineering and Applied Electronics,Beijing Universityof Technology,Beijing 100124,China)
出处
《计算机仿真》
北大核心
2022年第2期332-336,共5页
Computer Simulation
基金
国家自然科学基金资助项目(51775005
51675009)。
